Download 6.3 Power Strip Module - UCF EECS

Section1: Introduction/Definition
1.1 Executive Summary
At the Worlds Fair in 1893 Nicola Telsa captivated the United States when he lit
up the streets of Chicago with the first public demonstration of alternating current.
Since then the entire world had become dependent on AC power and it was
without exaggeration to say we would not be a technology driven society without
it. Electricity drove the industrial revolution just as it drives the economy today
and if all electric power were to fail suddenly the consequences would be
dramatic and life changing for all of us. If you take one minute to reflect on your
daily life and everything that it consists of it was nearly impossible to find
something that doesn’t use electricity or took electricity to make it. Electricity was
one of our most precious and useful resources and we must do everything we
could to safeguard it and make power failures as rare and isolated as possible.
It was the primary goal of this project to ensure that homeowners are unaffected
by utility company power outages and that they are able to continue their day to
day activities without interruption or inconvenience. With an upfront initial
investment a homeowner could easily make their home immune to power failures
and time spent in the dark. This could be done with the purchase of a home
generator from any home improvement store and the components designed in
our project. There are many sizes of generators and like anything else the bigger
and more powerful the more expensive. It was up to the homeowner to decide
what was in their budget and how much of their home’s electric circuitry they
want to use during normal power failure. One of the best features of our project
was it gave the homeowner the freedom to choose any size generator for their
house while not having to change any aspect of our design.
In our project we designed a system used to monitor power during normal and
emergency power operations to help save the homeowner money on their
monthly power bill as well as maintaining a safely operating system. The project
consisted of three key components: a power strip similar to the one you see
computer equipment usually plug into, a device for measuring home generator
power consumption, and also a touch screen monitor used for the user interface.
The power strip contained voltage and current sensing circuits along with a
microcontroller to calculate power consumption and quality. This processed
information was sent wirelessly via RF communication back to an embedded
touch screen microcontroller for display to the user. The main intent was for the
homeowner to have multiple power strips and together they report back to the
touch screen giving power consumption information for almost the entire house.
The device used for measuring the generators power consumption contained
essentially the same components as the power strip with minor modification, but
looked a little different to fit the output configuration of the generator. The touch
1
screen monitor would include a microcontroller and wireless RF communications
for receiving and transmitting data.
In order for the user to interact with the system a touch screen was used. This
touch screen had 3 inputs the user could manually enter. The first allowed the
user to manually add additional power strips to the network. The second input
also allowed the user to individually turn power strips on and off. Last the touch
screen also allowed the user to input the KW rating of their home generator to be
monitored as part of the system.
During times of receiving normal utility company power the system work as a
basic power monitoring system. Any device plugged into the power strip had its
power consumption and cost calculated and displayed on the touch screen. The
feature aided the homeowner on monthly power bill reduction and just overall
reduction in wasted power consumption. When normal utility power was lost and
the home started operating on backup generator power the touch screen
monitored power generation from the generator and power consumption from the
home making sure of acceptable operating conditions. The touchscreen kept
track of the average power usage of each device plugged into the power strip
and aided the user in deciding how many appliances their home generator could
handle at once. The whole point of monitoring the system in emergency power
mode was to ensure generator failure due to overload was not an issue.
Many homeowners lose power for days or weeks each year, due to hurricanes
and other natural disasters, which was why we believe something like this could
be so useful if it were available. Not only did this design aid homeowners in time
of distress, but it could also help them become more responsible with their bills
and make them a little more environmentally friendly.
1.2 Project Description
The intentions of our project were to design a home energy management system
that would be beneficial to the homeowner during normal power operating
conditions and also during times of utility power loss where the home generator
kicked on and took over. We designed power strip modules that resembled
those that you would plug all of your computer and office equipment into at your
desk. These power strip modules had voltage and current sensing circuits to
calculate power consumption on individual appliances. This information was
then sent back to a controller wirelessly that was then further processed to
determine things such as instantaneous power consumption, daily power
consumption, and total power consumption since last power bill. The controller
was then communicated to a LCD touch screen monitor that gave the user
access to the system. The user needed access to the system to add additional
power strip modules to the system. The power strip modules also contained
contactor circuits to allow the user to remotely turn on and off appliance through
the LCD touch screen monitor.
2
The concept behind the entire project was as follows. During normal power
operating conditions the homeowner may have used the LCD touch screen
monitor to observe total power consumption, rate of power consumption,
individual appliance power costs, and also projected monthly power bills. The
user also had the ability to remotely switch on and off individual receptacles in
the power strip modules to meet their monthly power consumption goals.
An affordable generator for home electrical power backup was typically unable to
support the full load of the house. Thus, during emergency power operation
(i.e. the whole house running on the store-bought generator) the user had the
same options as they did during normal power operating conditions with the
addition of the following features. Our project used current and potential
transformers to calculate how much of the total available power the generator
was producing and wirelessly sent this information back to the main controller for
processing. During the initial transfer between normal utility power and generator
power, the controller automatically switched off all loads connected to the power
strip modules. From this point the user then used the LCD touch screen monitor
to decide what appliances to power on. The controller then looked at how much
power was currently being used by the homeowner compared with how much
power was available from the generator and signaled the homeowner if they
were about to overload the generator, which resulted in a shut down. This
feature allowed the homeowner to use any appliance connected to one of our
power strip modules while insuring the home generator was not overloaded.
The controller was also programed with a safety feature that automatically shut
off circuits during overload to maintain power in case the homeowner did not
respond quickly enough.
1.3 Motivation
The motivation our group had for choosing this project arose from something that
affects those of us living in Central Florida closely, but its applications and uses
were not limited to this region. Each year, powerful tropical storms threaten the
entire Eastern coast of the United States, and a large number of these hurricanes
find themselves in Florida. These storms could leave hundreds of thousands of
people in Central Florida alone without power, access to roads, and the ability to
communicate over the internet for extended periods of time. In 2004, hurricanes
Charley, Frances, and Jeanne arrived one after another in quick succession,
leaving much of the region without power for weeks. Often, there was too high a
demand on the limited number of workers available from the power companies to
restore power in a timely manner. Thus, many homeowners decided to invest in
generators to help power their homes during these times. However, most
average people do not have the time to learn how exactly a generator works, and
may continue with their normal routines once the alternate power source was
hooked up and running. Unfortunately, this situation could lead to an overexertion
and failure of the generator. For these reasons, we choose to design a home
energy system which completely automated anything and everything about the
3
generator to avoid this possible situation, while also giving the user an
appropriate amount of control over the power supplied to their home. Not only did
this protect the homeowner’s investment in the generator, but the system kicked
into place as soon as the main source of power was lost (i.e. the incoming power
from the utility company), thus potentially saving the user weeks of time without
power waiting for it to be restored.
1.4 Goals and Objectives
The intent of our project was to design a home energy management system. This
system would be versatile in that was beneficial to the homeowner during both
normal power operating conditions as well as during times of utility power. Our
definition of “normal power operating conditions” was when the homeowner was
connected to a power source from the utility company. “Utility power” was defined
by the entire house running off of a generator which alone did not provide an
amount of power equal to what would be used from the power company on a
daily basis. In fact, the utility power should have been an amount that would have
severely limited the abilities of the homeowner to power the appliances in their
home. This was an important distinction because much of our project was
focused on the sometimes difficult choices that would need to be made during
times of power outage. Often, the average homeowner was unaware of the many
details and tremendous amount of power that go into allowing their lives to run as
normal. Questions like, “Can I run my hot water heater on a generator?”, “How
much power do I need to keep the refrigerator on?”, or “What are my limitations
for powering my computer/television/modem during power failure?” were
simplified and answered in our design, while still giving the end-user as much
control as possible.
The main design of our project began with the loss of normal power, at which
point the generator already connected to the user’s home turned on in response
to a signal from a microprocessor. That microprocessor was constantly
monitoring the current coming from the main, “normal”, power source; as soon as
the current disappears, a signal was sent to the generator to turn on and take
over power for the house. Our design was a power strip module that resembled
the surge protector power strips which could be used at one’s computer desk to
plug in your computer and its peripherals. It had two or three outlets to plug in
any standard utilities. These power strip modules had voltage and current
sensing circuits which would be constantly taking measurements related to
power consumption. The power strip modules also contained contactor circuits to
allow the user to remotely turn on and off the appliances through the LCD touch
screen monitor. This information (current and voltage readings) was then
processed at a microcontroller (embedded in the power strip module), and then
transmitted wirelessly to the main microprocessor in the LCD.
The wireless communications in our system were done using a single protocol,
which was yet to be determined. Shown below in Figure 1, each wireless device
4
was a transceiver (i.e. had the ability to transmit and receive data); they did not
communicate with each other, but only to and from the central microprocessor (in
the LCD). The wireless transceivers were used over a short range in our
prototype; this could have been extended to the end-user as well, assuming the
main processor in the end product was located near the center of the house. This
aspect would allow for lower power usage, as well as lower costs for the
transceivers.
Figure 1: Wireless communications block diagram
The LCD contained the main “hub” microprocessor which collected data from the
power strip modules and was where the bulk of the calculations and decisions
were made. This display was the interface between the user and the software
which controlled whether or not outlets in the power strip module were supplied
with power. During normal operating conditions, the user was able to control
which outlets were turned on (supplied with power) and also have the ability to
turn off any or all of the outlets. There was also a section on the LCD which
provided the user with pertinent data about their power consumption. In this
section, an emphasis was placed on “useful” data; that is, we didn’t wish to give
overwhelming amounts of graphs and numbers which may have required
knowledge of math to interpret and understand. The LCD showed information in
as simple a form as possible, while still providing important, useful data. The data
shown was facts such as:





Instantaneous power consumption
Individual appliance power costs
Daily power consumption
Current rate of power consumption
Total power used since the last power bill
5
An emphasis was placed on showing as many dollar amounts as possible, so the
user could more easily relate to the information on the screen. Abstract figures,
graphs, and even most other data was shown for the purposes of the
demonstrations, and on a limited basis at that. This process was intended to help
the user to more accurately monitor control their power usage, in order to meet
their monthly power consumption goals.
1.5 Requirements and Specifications
The design utilized an automatic transfer switch to transfer power to a secondary
power supply in case the primary source fails. The automatic transfer switch was
a third party device. It only needs to be tested for functionality. For this project
the use of a generator to provide secondary power in case of an outage was
intended. After power had been transferred the devices were able to read the
load from the generator and communicate this information via wireless signal to a
main microcontroller.
Other outlets in the home also had power monitoring features. They had the
ability to both transmit and receive instructions or data. Each outlet was required
to read the current that was being drawn by that specific outlet. That information
was then relayed to the main microcontroller which summed all of the power
being drawn from the home at that time. If too much power was being drawn
with respect to the generator an interrupt signal was sent to any of the
microcontrollers that were in control of an appliance. These microcontrollers
stopped the use of the appliance immediately upon request.
Each microcontroller measured a specific current. The accuracy of these
currents was precise enough to prevent damage to any appliances or the
generator. We set our initial goal to reach a precision of 1mA. The main unit
received power reading information from all the other microcontrollers every
second. The main unit provided absolute instructions to all other microcontrollers.
It must have the ability to stop and resume power consumption of any outlet in
the home. These criteria were set by the user through an LCD input screen. In
case of an overload of the secondary power source appliances shut down with
respect to their priority.
Outlets were sorted by priority based on user input. Each outlet priority level had
a number ranging from 1-10 with 10 being the most important. In the event two
outlets are given the same number the user was asked which appliance was
more important. Here the user could decide if they would rather have warm
water or air conditioning. All of the priority level inputs were given to the system
by the user through the LCD panel. In case no priority was set the appliance
draining the largest amount of electricity was shut down.
The power strips were able to achieve these specifications as part of our design:
6









Send data every 5 seconds
Enter sleep mode for 5 seconds in between packet transmissions.
Measure currents up to 15 amps
Measure voltages of 240V
Up to 2 appliances could be connected to any power strip
Have a line of sight range of at least 100 feet in an indoor environment
Have a range of at least 50 feet between two rooms
Power readings were measured within an error margin of 5%.
Had an ADC that was 10 bits wide for 1024 steps of resolution
The Main control station consisted of a controller that received information from
all the power strips and displays this information on the user controlled LCD.
This station was able to send and receive information from the LCD. If the LCD
relayed information to shut down a specific appliance an interrupt signal was sent
to relay that information.
The specifications for this unit were as follows:





Receive data from 3 different power strips 5 seconds
Receive data from power metering device two times per second
Communicate this data to the LCD module 3 seconds
Had an ADC that was 10 bits wide for 1024 steps of resolution
Receive and transmit information with an accuracy of 3 decimal places
with the local microcontroller.
The LCD was driven by its own processor and was able to communicate with the
microcontrollers. Through the LCD screen the user had the ability to read power
readings from any appliance in his home as well as the readings from all of his
appliances at once. The user was able to set priority levels for each of these
outlets. In addition the user had control for turning on and off each outlet
individually. If the user was trying to activate an outlet that would overload the
generator then they were given a warning. The LCD and the main control station
were in constant communication.
The LCD screen was required to do the following:

Receive and transmit information with an accuracy of 3 decimal places
with the main microcontroller.
While an outlet was not being used it entered a sleep mode to preserve the
battery life of that specific unit. The microcontroller remained in sleep mode as
long as there was no signal telling it to turn on. Once the microcontroller
received the signal to turn on it returned to normal operation.
The last power metering device was the one reading the power information being
delivered by the generator. This device relayed the information to the main
microcontroller. All of the overloading decisions of this project were dependent
7
upon this part. This device was able to read the power load on a generator by
reading voltage and current information.
The generator power metering device was required to do the following:





Read power supply information from the generator twice a second
Sent the power consumption information to the main microcontroller twice
a second
Must have an accuracy of 3 decimal places
Measure currents up to 30 amps
Measure voltages of 240V
1.6 Roles and Responsibilities
Our design group consisted of four senior electrical engineering students. The
members included Chris Diller, Christian Aranha, Kurt Riecken, and Arman
Murat. We decided to split the research and design into four main parts where
each group member was given an equal amount of work. Each of us were
working on separate parts of the project, but at each group meeting we explaind
to each group member what we had accomplished and how it worked. We
believed it was very important that each group member was completely
competent in understanding the entire design of the project.
Chris Diller had taken on the role of researching and designing the user interface
of the system. Chris was responsible for researching the different kinds of touch
screen monitors available today and the pros and cons each. He was also in
charge of selecting the main microcontroller to power the touch screen monitor
and wireless RF communications between power strip modules. Chris was
designing appearance of the GUI interface and was also writing the code to
implement it. Last he also wrote the code for the microcontroller receiving and
transmitting data from the power strip modules and made sure that information
was displayed neatly and accurately to the touch screen monitor.
Kurt Riecken had taken on the role of researching the microcontroller needed for
the power measurement and calculations. Kurt was determining whether a
standard microcontroller such as the Aurduino was needed or maybe something
more application specific. He was also in charge of processing the signal
received from the voltage and current sensor and making the appropriate
calculations for power measurements. He then wrote the code for the
microcontroller to receive and transmit data from the touch screen display Chris
was designing.
Arman Murat took on the role of researching and designing the circuits and
sensors needed for power measurement. Arman was researching the available
technologies for current and voltage monitoring and determining what fit the
project best. Arman and Kurt were working closely together to ensure
8
compatibility of sensors and microcontrollers and any additional circuits needed
to connect the two and make them operational.
Christian Aranha took on the role of researching and designing the wireless
communications between the major components in the project. Christian was
researching the latest technologies in wireless communications and determined
what technology worked best for our project. Christian was also in charge of
researching the analog to digital and digital to analog conversions needed for
transmitting and receiving the data from the microcontrollers.
We also
considered creating our own wireless protocol for transmitting our data back to
the touch screen and Christian also did the research on this.
Additionally we had a small amount of our project that required minimal design
but was required for demonstrating purposes to the class. This part of the project
included researching automatic transfer switches, circuit breakers, and
receptacle circuits. Team members contributed their ideas on how the project
should be demonstrated and what it took to do so.
Section 2: Research
2.1 Similar Projects and Commercial Product
After the decision was made about what our project would be as a part of our
research we went through some other projects presented by other students. In
our research through different colleges senior design projects, we all agreed that
UCF projects were better in countless aspects and worth to review in the “similar
projects” section in our paper.
The first project that was similar to ours was from Fall06-Spring07 done by Group
24 with the project name H.E.M.S. (Home Energy Monitoring System). Just like
our project’s goal this group intended to save money to the user by conserving
power. Their system basically had the same features just like ours such as
measuring the power consumption at the current moment or overall through the
whole month, transmission of this information wirelessly to the main processor
and the user interface, LCD display for user interaction. Although both projects
were trying to achieve the same goal, our senior design project had additional
features on top of the mentioned above. One difference they had in their design
was installing the power measurement components to the gate where power
entered the house; the circuit panel. This made many things easier and less
complicated since the whole power consumption could be read from the circuit
panel. Since our projects had different features (i.e. in case of a power shortage
our designed system automatically started the generator but not only started it
but distributed the power produced by it according to the importance level of the
outlets) this way limited generator power would be used efficiently. In order to
achieve this system we chose to install the power measuring components into
9
individual power strips which would go to every individual outlet in the sample
demonstration house. In our design project we were going to have the ability to
turn on and off several outlets from the LCD control panel. If we were to
summarize the differences between two projects our design would be able to do
the home energy monitoring and in addition to that we would be able to do
emergency power monitoring and management as well.
The second project that was somewhat similar to our project was from Fall06Spring07 done by Group 3 with the project name Airpax Smart House. This
project was not about power management but it had some features that save
power such as motion sensors that turn lights on and off when a person enters a
room or leaves. Similar features they had were the wireless communication
methods and LCD display control but basically this project was a smart house
computer with a couple of power saving features but not a power management
system.
The third project was another power management project Fall07-Spring08 done
by Group 17. The project’s name was Occulight. The group members’ goal was
to reduce power consumption by placing motion sensors into each room and
adjusting the power consumption in each room by their occupancy. This project
was very much like the previous project but addition to turning lights on and off
this project want to take it a step further and be able to switch power outlets on
and off according to the rooms occupancy and power outlets importance level.
Their choice of communication was wireless the same as our projects but again
our projects had different features such as in case of a power shortage our
designed system would automatically kick the generator in but not only leave with
that but distribute the power produced by it according to the importance level of
the outlets in order not to overwhelm the generator.
Next project was again somewhat similar to our project was from Fall07-Spring08
done by Group 18 with the project name Detect All 3000. This project was not
about power management, it was an emergency fire and hazardous gas
detection system. The similar characteristic of this project to ours were its
sensor’s interaction with the main board and its LCD control set-up.
The fifth project that caught our attention as a similar project from UCF senior
design data base was from Spring08-Summer08 done by Group 1 and the
project’s name was Power Monitoring System. Similar to our project’s goal this
group’s project was to save money to the user by power conservation. The
design project basically had the same features as ours such as measuring the
power consumption at the current moment or overall through the whole month,
transmission of this information wirelessly to the main processor and the user
interface and LCD display for user interaction. There were many differences
between the two projects as well. Our senior design project had additional
features than theirs. Our project had an emergency feature: in case of a power
shortage our designed system would automatically start the generator but not
10
only started it but distributed the power produced by it according to the
importance level of the outlets, this way limited generator power wouldn’t be
overwhelmed. In our design project we had have the ability to turn on and off
several outlets from the LCD control panel; the outlets were at the same time
also power measuring units. While we had additional features, this group’s
design had different features than ours as well. Their profiling system for major
appliances in order to turn them off for the time the house hold was empty
justified itself as a big energy saver.
The next project which was again somewhat similar to our project was from
Spring10-Summer10 done by Group 2, with the project name Wireless Power
Meter. This project was almost the exact same as ours again. This group did
have a power strip power measurement device and a wireless protocol to
transfer data to the main control device. They did have a set accuracy
requirement of at least 5%. Again they lacked the emergency power
management feature that our project had where smart generator power
management was achieved in case of a power shortage occurring.
Another project similar to ours was from Summer07-Spring07 done by Group 6:
the project’s name was Taj’s House. This project again tried to save energy and
cut the power bill for the user but the energy management was achieved not
automatically but remotely through a Bluetooth device. The idea was to be able
to turn on and off the light switches and thermostat using a Bluetooth device this
way if the user wouldn’t be home for a long time energy consumption was
reduced by turning these switches off. [23]
Our last example came from Fall10-Spring11 group 13. The project name was
Home Energy Management System. Their project was very similar to ours and a
couple of other projects above where they had a power strip power measurement
device and a wireless protocol to transfer data to the main control LCD device.
Our emergency power management feature separated our project from theirs but
everything else was very much alike except our choice of microcontroller and
wireless communication device choice.
After the review of some similar projects we wanted to review some similar
products that did the same thing as well as available for purchase in the market.
The first one that drew attention was called Kill-a-Watt produced by P3
International. There were a variety of products developed and produced by the
corporation and basically these components were more or less what we wanted
to achieve with our project except the smart emergency power solution feature
we have in addition to this power saving and monitoring system. Below is the
generic Kill a Watt power strip shown and the other model that was more similar
to our project since there was wireless communication to an LCD screen where
output was displayed and user interaction was possible. Unlike our design, there
was no switching on and off feature of the power strip available from this LCD but
11
an impressive 0.2% accuracy was achieved. In figure 2 below the wireless
module and the display screen are shown. [23]
Figure 2: P4225 Kill A Watt®
Permission requested from www.p3international.com
2.2 Home Electric Service
The two primary forms of electrical distribution here in the United States were
single-phase and three-phase power. Single-phase power was used for all
residential homes and very small business offices. Three-phase power was used
in commercial and industrially applications and was not considered in this project,
as our design would have had to be completely modified to account for it.
Single-phase power was a single alternating power source that was sent out on a
conductor that measured a voltage potential between it and a ground wire. This
was a two-wire system where current flowed from the source through the load
and then back through the ground wire. Because it was an alternating source
each 180 degrees the current reversed direction and headed in the other
direction. In this type of system, only one voltage could be obtained for utilization
and that voltage was dependent on the physical characteristics of the generator.
12
A residential home had three conductors coming to it from the power company.
Two of them were “hot” conductors and the other was a “neutral” conductor. The
hot conductors measured 120 volts between them and the neutral conductor and
they measured 240 volts between each other. How could this be possible if the
service were a single-phase system? The power company really was supplying
the house with a single-phase system, but the transformer supplying the house
was playing a little trick to do so. Figure 3 shows a diagram of the utility
transformer and the conductors on the input and the output.
Figure 3: Home Utility Power
As you could see from the figure the power company’s wires came in on the left
and only included 2 wires. A voltage of 7200 V between the hot and ground
conductor was most commonly used on the low power poles located in
residential areas. On the load side of the transfer were the “A” and “B” phases,
which were by definition the outputs of the conventional transformer. In the
middle of the secondary winding of the transformer was a center tap and created
a point of equipotential for the “A” and “B” phases. Through simple circuit
analysis it could be shown that the potential between phases “A” and “B” was
240 volts while the potential between phases “A” or “B” and the neutral conductor
was 120 volts.
The reason for needing different voltages in a house was to serve the different
appliances in it. Large equipment such as ranges, air conditioners, clothes
driers, and hot water heaters were all 240-volt equipment because it allowed for
them to remain a reasonable size and have a more efficient design. Every
appliance surely could be designed to operate at 120 volts, but appliances
designed at higher voltages could operate with a low current and therefore
reducing the 𝐼 2 𝑅 losses of the appliance.
13
The first prototype of this project was designed around the appliances that
operated at 120 volts and if time and budget permitted, the design could have
easily incorporated the appliances that operate at 240 volts with minimum
modifications.
2.3 LCD Touch screen Displays
A touch screen device was chosen for user interface for the simplicity of use for
the user. This was the only part of our project with which the user interfaced so
to ensure user satisfaction we wanted to choose a user-friendly interface that
anyone from a child to an elderly person could operate with no training making
sure not to aggravate them. It was of the utmost importance of this project that
the user would embrace the power and usefulness of the product because of
their understanding and eagerness to use it.
There were many types of different touch screens on the market, but the 4 most
commonly encountered and used included resistive, surface wave, capacitive,
and infrared. Each type of screen had its advantages and disadvantages as well
as what application for which it was best suited. For this project we considered
the fact that the touch screen would be located in a household and mounted on a
wall near the thermostat. We also considered that this touch screen would be in
use for many years and would have many dirty fingers touching the screen
throughout the years. It was therefore important that during the final selection of
a touch screen for this project that we selected one that would withstand the test
of time and that the screen would have some kind of protectant to combat greasy
smear from dirty fingers or that it would be easily cleanable.
Our touchscreen controller also incorporated an embedded microcontroller used
for receiving and processing data from the various power strip modules
throughout the house. When selecting the embedded microcontroller we
considered the rate of transmissions the controller could output and receive, the
amount of memory available for programs and data, and also the computer
language the embedded controller used for programing. We planned on being
able to store one month’s worth of data for each power strip module to calculate
totals, power consumption, and monthly costs of individual appliances. The
embedded microcontroller would only require a few analog and digital I/O
because a wireless transceiver would be the only device connected to it. The
rest of the design lay within the power strip modules and it was here where all
power measurements would be taken and transmitted back to the touch screen
controller.
In the next few sections a brief overview of each of the four types of screens
were explained. Through comparison of the advantages and disadvantages of
each of the types of touch screens we were able to decide which satisfied the
requirements of this project the best. Even though we may have found a touch
14
screen that fit our application perfectly, another may have been selected due to
the cost constraints of this project.
2.3.1 Resistive Touch Screens
Resistive touch screens were some of the most cost effective and widely used
touch screens on the market. They were extremely durable and were less prone
to contaminates than acoustic wave touchscreens. Resistive touchscreens were
also less sensitive to the effects of scratches like capacitive touchscreens were.
The anatomy of a touchscreen was very simple and was one of the major factors
of their durability. The monitor was made up of two sheets, which were
separated by insulating dots. The top layer was flexible and was what the user
pressed their finger against while the bottom layer was rigid and was what gave
the display its firmness. Each of the sheets’ inside layer was coated with a
transparent metal-oxide, which in turn gave the sheets their electrical
characteristics. A voltage was then applied to the layers creating a gradient
across them. When a user pressed anywhere on the screen the two layers made
a contact at the point closing the circuit allowing the coordinates to be located.
Figure 4 shown below represented the structure of the resistive touchscreen
monitor as well as a list of its various components and where they were located.
[3]
Figure 4: Resistive Touch Screen
Permission requested from Fast Point Technologies
1.
2.
3.
4.
5.
6.
Polyester Film
Upper Resistive Circuit Layer
Conductive ITO (Transparent Metal Coating)
Lower Resistive Circuit Layer
Insulating Dots
Glass/Acrylic Substrate
15
7.
Touching the overlay surface causes the (2) Upper Resistive Circuit Layer
to contact the (4) Lower Resistive Circuit Layer, producing a circuit switch
from the activated area. The touch screen controller gets the alternating
voltages between the (7) two circuit layers and converts them into the
digital X and Y coordinates of the activated area.
2.3.2 Surface Wave Touch Screens
Surface wave touch screens were devices that utilized ultrasonic waves to
process inputs from the screen. The display in figure 5 had a glass layer that had
transducers on both sides used as receivers and transmitters. These transducers
were placed on the X and Y axes and were used to create a coordinate system.
An ultrasonic wave was generated and sent across from the transmitter to the
receiver. The screen’s glass layer had reflectors on it, that when pressed, the
energy of the wave created was reduced and the screen registered a hit giving its
location. A negative characteristic of the surface wave touchscreen was it’s
influence to bad performance from external forces such as dust, water, loud
sounds, etc. Anything that could distort the acoustic wave would potentially
make the device fail. [6]
Figure 5: Surface Wave Touch Screen
Permission requested from itechcompany.com
16
2.3.3 Capacitive Touch Screens
Another type of touch screen was the capacitive touch screen. There were
several different categories in which you could classify a capacitive touch screen
and they included surface, projected, mutual, and self-capacitive. Each of these
technologies differed from the others, but the fundamentals of how they worked
remained the same for all of them. A capacitive touch screen consisted of an
insulating material such as glass, which was coated with a transparent conductor
like indium tin oxide (ITO). A human finger was also an electrical conductor and
when that finger pressed against the screen a change in capacitance between
the two charges could be calculated and when sent back to a processor this
could pinpoint the location of where the finger tapped the screen. Capacitive
touch screens as seen in figure 6 had a considerable disadvantage to resistive
touch screens in the sense that they must have a conductor such as metal or a
human finger come in direct contact with them to select a location on the screen.
This was an undesirable trait for consumer electronic devices where in cold
temperatures a user might be wearing gloves or a woman with long fingernails
might have a hard time getting the screen to operate correctly. Capacitive
screens were in general more responsive than resistive screens, but were more
costly and were impractical for this project. [
Figure 6: Capacitive Touch Screen
Permission requested from eetimes.com
17
2.3.4 Infrared Touch Screens
The infrared touchscreens used a simple design approach to deliver great
accuracy and response. The screen used a grid of infrared LED’s and photo
detector pairs in an X and Y coordinate system around the outer edge of the
screen. These photo detectors were used to detect interruptions of the LED’s
resulting from a user pressing on the screen. The accuracy of location was
dependent on the number of LED’s and photo detectors along the X-axis and Yaxis. The more pairs the greater the accuracy. When the screen was pressed a
photo detector triggered a hit on the two axes thus giving the coordinate location
of the hit. Figure 7 below illustrated the yellow bars on the left and bottom as the
transceivers and the orange bars on the top and right as the receivers. You
could see how the distortion of someone’s finger pressing against the screen
would allow the location to be calculated. Infrared touchscreen monitors were
very durable and were used for industrial and military application so we knew
they would uphold to the standards of our project. However, the infrared
touchscreens were rather pricey and therefore would not be applicable to the
cost constraints of this project.
Figure 7: Infrared Touch Screen
Permission requested from ATouch Technologies
2.4 Power Monitoring Circuits/Instrumentation
Since the main goals of the project were to measure power and, if necessary,
restrict the usage of this power, the most imperative part of the project was to
measure the power correctly. This correct measurement of power was going to
reflect to the correct outcome and calculation of the energy cost. Appropriate
energy distribution happened to be another feature of our project. In case of a
power shortage the generator should kick in and limited power produced by the
18
generator would be distributed according to the power consumption of critically
important outlets without causing any overage of the generator output.
Our first question was which power we were measuring. Energy was distributed
in AC form for the reason of this form traveling longer distances and it was less
complicated to produce the energy in this form rather than brushless rectifier
generators. After the production of this energy was transferred with high voltage
lines with the least current possible to reduce the energy loss, voltage dropped
down in transfer locations which were big transformers and finally were received
in a regular house outlet 120 VAC which varied with 12 to 15 amps of max
current drown. In a regular home there were different appliances with different
uses most of which were resistive loads since they were used somehow to
produce heat. Besides resistive loads we had inductive and capacitive loads as
well which had outcomes on the magnetic field when connected to the current.
The outcomes of these two different kinds of loads gave different types of power
which were “Real Power” and “Reactive Power”. Our concern was with the Real
power since that was the one that we were charged for so our calculations and
measurements would be based on Real Power.
The types of power could be listed as Real power which was measured in Watts (
P = V * I * cosƟ ), Reactive power which was measured in VAR’s ( Q = V * I
*sinƟ ) and Apparent power which was measured in VA ( S = P + jQ ). Since the
power company charged us only on Real power, our power measuring circuit had
to be measuring the real power. This brought up the importance of power factor
in other words cosine of “Theta”. This was the angle between both the voltage
and current RMS values and between the Apparent power and Real power.
According to the previously mentioned formula “P = V * I * cosƟ” where Theta
was the phase angle all we needed to do was multiply the values we got from the
measuring circuit with cosine Theta and we found the power that we were going
to be charged for which shouldn’t be a great deal to achieve. The real problem
here was how we calculated Theta from the voltage and current measuring
circuits. Therefore, to measure the “Theta” or the power factor which was cosine
of theta all we had to do was read the time difference between zero crossings of
the waveform of both voltage and current waves. This difference was
proportional to the whole cycle period. If we were to accept the whole cycle as
360°; comparison of the time difference between zero crossings of the waveform
of both voltage and current waves divided by the period gave the ratio of phase
angle Ɵ divided by 360°. Of course all these calculations would be done by our
microcontroller “MSP 430”. The important point here was the implementation of
the current and voltage measuring circuits to the microcontroller since there was
an allowed range of inputs to MSP 430. This was an easy problem to solve by
implementing a voltage divider circuit, to both current and voltage measuring
components, to reduce the voltage to make it compatible with the
microcontrollers input values. The reduced factor in the voltage divider was
important since later on in the representation of the power this factor was
included otherwise the power consumption would have turned out wrong. [2]
19
After all we discussed about how we were going to measure the power it was
time to cover the components we would use for the actual measuring of voltage
and the current. Among numerous components that were suitable for the job we
would make an educated decision and try to pick the one with the best accuracy,
price and circuit compatibility. If this assignment was given in our first circuits
class we would have came up with a solution that involves ohms law.
If we were to attach a resistor to the conducting wire where we wanted to
measure the current and measure the voltage drop over the resistor we would
have known the current. The real life application of this procedure among current
measuring methods; the simplest one was the current shunt shown in figure 8.
Figure 8: Current Shunt
Although we briefly talked about the current shunt here we would not consider it
to be in our circuit. This was for several reasons. First one was the heat problem.
A resistor when contacted a current in a conducting wire produced voltage and
this means at the same time there was a power dissipation. The power in
resistors was in the form of heat. In measurement units in watts, 𝑊 = 𝑅 ∗ 𝑖 2
which could be interpreted as the more current the more heat. Heat was
something we didn’t want in our circuit since it could harm everything around it
starting from the shunt itself. This brought the second reason why we couldn’t
use the current shunt in our circuit. As mentioned before the project’s main goal
was to measure power with accuracy. Heat would tend to change the resistance
so the measurements wouldn’t be as accurate with the rise of the current since
the resistance would be a different value. Although shunts with a dependable
resistance drifts were available these were way too pricy and still not the best
choice since the size of these shunts were another inconvenience. This brought
us to our real evaluation of current measuring components:
2.4.1 Power Relays
Our system design was all about managing power and this required for us to be
able to turn on/off an outlet, which consumed too much power. In order to this we
need a switch, which we could control remotely and this switch was the power
relay. The relay in figure 9 was an electromagnet which moved an armature part
20
attached to a spring. When the electromagnet was powered the magnetized
metal part attached to the armature moves and either connects or disconnects
the circuit according to the design or what was tried to be achieved. When the
power for the electromagnet was turned off spring did the mechanical job to do
the opposite of what was achieved by the electromagnet.
Figure 9: Relay
In this initial research phase we looked at many different applications of relays
and saw that we could initially group them into two major groups;
electromechanical and solid state. The basic principal of both applications were
the same but we saw that solid state relays would be way more beneficial for our
cause. There were many reasons for this selection that should be mentioned. In
the electromechanical version one problem might be caused by the use of direct
current and the spike problem. A voltage spike was dangerous for a
semiconductor and it was a big probability for the coil to have a voltage spike
when the power was switched off and this problem was eliminated with a diode
where this time there might have been different issues with AC power.
Alternatively, a parallel capacitor and a resistor could solve this same problem
talked above. Generally speaking, an electromechanical relay such as a latching
relay in figure 10 would be a perfect candidate for our project since it would serve
the project’s energy saving purposes since its energy consumption was almost
zero. The way a latching relay worked was a little different than the conventional
simple relay described above. Here there were two coils instead of one and a
permanent magnet that kept the armature arm remaining in the position it was
located after the pulse, so there was no continuous current needed to keep the
electromagnet charged to keep the armature in the same position. Individual
pulses to each coil would give either one or zero so that the switch operation
would be achieved. Another advantage of the latching relay would be its low
resistance values which would mean low power consumption and less heating.
Price-wise it was considerably cheaper compared to other candidates. Size was
not as bad neither which was very important for the design of the strip module we
were designing; though, compared to the solid state relay it was twice as big. Its
power handling capabilities could be phenomenal. For the “Contactor” type of
21
electromechanical power relays high loads such as 10 to 50 amperes could be
achieved but this type again came with a time delay in order to secure the
equipment. Beside the size comparison the real reason we decided not to
choose latching relay was the complicated design process. Electromechanical
relays needed separate wiring and a separate control algorithm.
Figure 10: Latching Relay
In our search for relay we didn’t even consider reed, mercury-wetted, polarized
and contactor relays. The first two; reed and mercury-wetted relays were almost
the same thing with the difference of the first one had its contacts in a vacuum
and the second dipped in mercury for the reason to stop the corrosion. They
were both too big and for small voltage usages. The polarized relay was too big
as well and it was used for high precision applications, which was not what we
needed for the relay part of our project. As mentioned before a contactor relay
was a high voltage switch. It could handle high amperages but, since a home
outlet had 12 to 15 amps and the noisy operation of this relay was a good reason
for us not to consider it.
After reviewing all the pros and cons of the electromechanical relays it was time
to take a look at SSR, in other words Solid State Relays. Since in solid state
relays displayed had no electromagnet that part may be replaced by a LED and
an Opto-TRIAC as shown in figure 11 below and everything worked flawlessly.
22
Figure 11: Photo Coupled SSR
Permission requested from http://www.pc-control.co.uk/relays.htm
Solid State Relays were semiconductor based switches and unlike electro
mechanical relays, they didn’t have moving parts which meant they didn’t need
maintenance unlike changing of the coil was had to be done in contactor relays.
This made them very reliable with their lifetime. As mentioned above, an
electromechanical relay, such as a latching relay would have been a perfect
candidate for our project since it served the project’s energy saving purposes
because its energy consumption was almost zero. Its low resistance values, not
only meant low power consumption but fewer heating problems as well.
Pricewise it was considerably cheaper compared to other candidate the SSR. So
what made Solid-State relay a better choice?
An LED Opto-TRIAC was the type that came into mind when an SSR was
mentioned, since it was the most popular of all. Instead of the electromagnet, a
light emitting diode (LED) turned on with the application of power to the general
purpose relay and shone through a gap where an Opto-TRIAC was placed on the
other side. The TRIAC was a small semiconductor switch which was very similar
to a diode. TRIAC would be used by itself and would be a very inexpensive
choice with good switching power consumption for resistive loads, but it was not
the same for reactive loads since it required additional circuitry for applications
with reactive loads. An Opto-TRIAC was almost the same component described
above but this time the response of it was against light flashing across the gap
coming from the LED. This response opened and closed the circuit that it
controlled. Solid State Relays were known for their quick responses and their
vulnerability to electric noises, high price, high series resistance however way
more sensitive output in terms of rating compared to the latched relay switch and
their dependable long life and having no moving parts made up for their
drawbacks.
23
Just like the electromechanical relays, SSR’s had different types as well. The
most common being the Photo-coupled SSR or as it was mentioned above as
LED Opto-TRIAC SSRs, followed by the isolated Reed-Relay coupled SSR
shown in figure 12 or in other words Hybrid SSR and Transformer-coupled SSR.
Talked about above reed relays were vacuumed relays and in this application the
signal usually came to the relay pre-amplified and through the relay trigger circuit
reacted with the TRIAC trysistor switch. This application generally provided
excellent isolation because of the vacuumed reed relay property.
Figure 12: Reed-Relay-Coupled SSR
Permission requested from http:// www.wiringdiagrams21.com
The last type of SSR was Transformer-Coupled SSR in figure 13. In this
application this time signal passed through a DC-AC converter to guarantee that
it was AC signal to be processed which was basically a transformer and this
primary excitation’s end product as the secondary excitation got to the TRIAC for
the turn on and off.
Figure 13: Transformer-Coupled SSR
Permission requested from http:// www.wiringdiagrams21.com
After all the initial research, we decided to go with the Photo-Coupled SSR. First
of all this decision was made against electromechanical relays since their sizes
were too big and the idea of moving parts rose questions about their reliability.
They were very quick in responding to any turn on and off input and there was no
24
internal arching that would harm any semiconductors. Another advantage we
considered was its PCB compatibility both because of its small size and better
technology. When powered with AC source, solid state relay had the benefit of
zero crossing switching which reduced noise in the as a result the circuit could
experience switching where the voltage crosses were zero.
2.4.2 Magneto Resistive Field Sensors
Shown below in figure 14 was a magneto resistive field sensor. It reminded us of
the adjustable resistors that we used in electronics lab. The principle was: with
the help of very thin ferromagnetic films, a magnetic field was created on the wire
that carried the current. This magnetic field was used to adjust the resistance by
tilting it in certain angles and changing the magnetic field intensity. The
resistance was the largest if the magnetization and the current were acting in the
same direction, and the adjustment could be minimized if the angle of
magnetization and the current had 90° in between. [22]
Figure 14: Magneto Resistive Field Sensors
Permission requested from http://www.directindustry.com
There were four magnetic field sensitive resistors in a sensor and these resistors
were designed to form a bridge. This was called the measuring bridge and the
way it worked was between the four magnetic field sensitive resistors any
detected magnetic field change was measured as a potential difference along
each side of the bridge. This magnetic field created might as well have been
coming from a wire and the intensity of the magnetic field created by the wire
could be measured this way and could be converted to another electric variable
in this separate circuit and be used to measure the intensity of the current. The
important thing here was the biasing of the sensor with a permanent magnet so
that the measurements could be accurate for the magnetic field; read, this way,
there was no need for saturation which would cause the precision of the
25
measurement to change. Usually the output voltage was a square wave voltage
which was easily convertible to digital with a Schmitt trigger. [22]
The way magnetic field sensitive resistors told the intensity of the magnetic field
was actually pretty simple. The full bridge formed from 4, or the half bridge
formed from 2, magneto resistors. Once put into a magnetic field they responded
to it. This response since they were on the opposite sides of the bridge and
placed inverted from each other was opposite from each other; in other words
while the first one’s resistivity increased the other ones resistivity decreased and
the potential difference created between the opposite sides of the bridge was the
measurement method of this change. Just like any other magnetic component
that functioned with magnetic fields the linearity of this component was not that
great and compensation with an aluminum conductor had to be implemented in
to the component, usually above the magnetic resistors, so that better linearity
was achieved. In hindsight this linearity was not that big of a deal after installing
this conductor since the current that was created in this conductor by the
magnetic field did the compensation job itself. This application at the same time
added to the convenience of this component since it compensated for the
temperature dependency along with its other advantages such as its good
accuracy and high sensitivity. At the same time it could take a lot of mechanical
stress, it was durable, it had fast response rate, had a decent size, could be used
in harsh conditions and the cost was reasonable compared to other types of
current sensors.
It was a tough competition between Magneto Resistive Field Sensor and Hall
Effect Current sensor since they were both very good options for our project’s
current measuring component. Against all the advantages of Magneto resistive
field sensor, the Hall effect current sensor had its own advantages such as: they
were logic capable, highly linear, non-mechanical structure, resistance against
heat, perfect to measure high currents but they wouldn’t break easily, were
insensitive to dust, vibration, humidity, cold and hot so their characteristics and
measuring sensitivity remained the same because of their well-sealed structure.
The downside of Magneto resistive sensor was its limited linear range,
temperature drift and its sensitivity to interfering magnetic fields from other
components compared to that of the Hall current transformer sensor. Once the
temperature characteristics were compared, the Hall Effect current sensor was
superior as well since magneto resistive current sensor could also have
temperature drift.
26
2.4.3 Hall Effect Current Sensors
The first discovery and where the definition of the Hall Effect came from was a
rectangular thin conductor, in this particular case it was a gold sheet with current
running through, got placed in a magnetic field perpendicular to the direction of
the current. The magnetic field created effects such that while the current was
uniform without it, with the magnetic field on, the concentration of current lost its
form and the distribution of the current lost its uniformity. This was called the Hall
Effect which resulted with a potential difference called the Hall voltage. The
important point here was the output voltage, in other words the Hall voltage was
proportional to the magnetic field and the current’s cross product and this was
the basic principle of the sensing circuit. Below in figures 15, 16, and 17 the
diagrams described how the Hall Effect worked. In the first representation, in
figure 15, there was a conducting plate shown. In our case this was the main wire
that the power strip was plugged into instead of the conducting plate that had
current running through.
Figure 15: Creation of Hall Effect with the Magnetic Field Representations
Permission requested from http://www.explainthatstuff.com
The blue dots were the electrons’ representation where our projects power
measuring strip drew from the main grid. Below though, the same conducting
plate was placed between the poles of a magnet. This at the same time meant
that the electrons passing through the conducting plate were under the influence
of the magnetic field as well. [21]
27
Figure 16: Creation of Hall Effect with the Magnetic Field Representations
Permission requested from http://www.explainthatstuff.com
The reason for this movement of the electrons was the Lorentz Force or in other
words any engineering student would call this effect as the good old “Right Hand
Rule”. Since the regular path of the electrons was interrupted and a deviation
occurred towards the bottom of the figure representation, now we could say that
the opposite sides of the conducting plate carried opposite charges (as
represented in the above figure). This created a potential difference and was
called the “Hall Voltage”. [21]
Figure 17: Creation of Hall Effect with the Magnetic Field Representations
Permission requested from http://www.explainthatstuff.com
28
Now, if we were to assume that we had more of those blue dots above in the
figure which represented electrons. This would mean there would be more
electrons gathered at the bottom of the figure and this would mean there was the
same amount of holes on the top of the representation and this would mean
there was more potential difference; in other words more “Hall Voltage” created
proportional to the intensity of the current passing through the conductor. [21]
So in this sense, developing a component where the magnetic field was facing
the wire from all directions or sometimes four directions with constant magnetic
flux would get the current exactly proportional with the Hall voltage. The more the
voltage the more the current principle with almost perfect measuring accuracy
would be achieved since magnetic field supplied from all directions would take
the stress factor out of the equation.
So what were the advantages of using a Hall Effect current sensor? There were
countless advantages, which varied from the safety to precision. Now let’s go
through these advantages, which led The Hall Effect Sensor to be our pick for
our project.









Hall Effect current sensor shown below in figure 18 could be isolated from
another high voltage in the same system. The real life applications of this
fact were usually the safety applications to measure high current and
voltage of systems that were even dangerous to get close to.
Hall Effect current sensors were insensitive to dust, vibration, humidity,
cold and hot so their characteristics and measuring sensitivity remained
the same thanks to their very well-sealed, closed structure.
Hall Effect current sensor’s other advantage would be considered to be
internal temperature absence since they didn’t get hot because of their
inductive structure but not resistive.
Hall Effect current sensors were convenient to use because of their nonmechanical structure same as their immunity against heat. So not only
they were perfect to measure high currents but they wouldn’t break easily
since there were no mechanical moving parts.
They did have a very long life because of the above reasons.
Hall Effect current sensor were more robust than others, others would
wear out by time because of arching
They were logic capable.
High speed and repeatability.
Highly linear.
29
Figure 18: Hall Effect Current Sensors
2.4.4 Eddy Current Sensors
Eddy Current sensors like in figure 19 were another magnetic field based
measurement component. Although it was called Eddy Current Sensor most of
its use was done for measuring distances. Because of its name we kept it among
the titles of our research material for the current sensors but despite its name the
real use of it were to measure distance, position, amplitude, alignment,
roundness, pitch and permeability of an object. Though, if the distance was kept
constant, the application could be used to measure the current as well. The basic
structure of the component was just a probe attached with a sensor that created
an alternating current at the tip of it. The alternating current created a magnetic
field naturally and this was the method Eddy current Sensor functioned, with
these magnetic fields. If we were to assume this sensor attached to the tip of the
probe as a coil with AC running through it, we would expect magnetic field
shooting out of it. If we were to get the probe close to a conducting material these
magnetic fields would have had an effect on the conducting material. One of
these effects was usually unwanted but in this case worked to the users’
advantage, Eddy Currents.
30
Figure 19: Eddy Current Sensors
Permission requested from /www.sensorcentral.com
The magnetic fields created by the eddy currents opposed the magnetic field
directed to the target conducting material. The sensor measured the interaction
of these magnetic fields. As mentioned above the current measuring application
would be achieved; the sensor could measure the change in magnetic in the
same way the electric fields and produced a voltage that was proportional to the
change in distance but if the distance was kept the same, this time the change in
the magnetic field intensity would give the change in current, since the more the
current more the magnetic field surrounding it proportionally existed.
The advantage of using an Eddy Current Sensor would be its tolerance to dirty
environments, having another material between the gap of measurement not
being a problem but since our power measuring strip was going to be closed and
would be for inside the house usage, environmental dirt nor having another
material in between the measurement gap would not be an issue for our
application. Despite the fact, using Eddy Current Sensor would actually cause
some problems. Since the main goals of the project was to measure power and
then if it was necessary restrict the usage of this power, the most imperative part
of the project was to measure the power correctly. This correct measurement of
31
power was going to reflect to the correct outcome and calculation of the energy
cost. A wrongful measurement would jeopardize the whole project and using an
Eddy Current Sensor would cause that. In the power measuring strip there would
be many components wired all together in a narrow space. More or less each of
these components would have a magnetic field or Eddy current that would
interrupt the measurement precision of the Eddy Current Sensor. Besides that
our other option Hall Effect Current sensor provided way more advantages for
our measuring circuit such as High speed and repeatability highly linearity, a very
long life and isolation from another high voltage in the same system so it wouldn’t
be affected by the other components just like Eddy Current sensor would. On top
of all these advantages against Eddy current sensor, Hall Effect current sensors
were insensitive to dust, vibration, humidity, cold and hot so their characteristics
and measuring sensitivity remained the same just like Eddy Current sensors.
2.4.5 Current Transformer Sensors
Current transformer Sensors were another application of power and current
measuring systems. The way the component worked was very simple and very
much like the basics of electric power production. We all knew that a conducting
wire that had current running through created a magnetic field surrounding the
wire. With the right hand rule we could find the direction of this magnetic field as
long as we know the direction of the current flowing. Current transformer sensors
were based on the principle to capture this magnetic field in another conductor
and use the magnetic field in it to measure the current running in the wire it
surrounds. The way to do this was to wrap another conducting wire around this
conducting material that surrounded the initial cable and this time created current
in this wire induced by the magnetic field running through the conducting
material. In figure 20 a current transformer was shown.
Figure 20: Current Transformer
Permission requested from electro-tech-online.com
32
This current was proportional to the original current which we wanted to measure
at first and the ratio was the number of windings that were wrapped around the
conducting material that had the magnetic field going through it.
There were three major types of Current transformers and the description above
may have gone for either Window, Bar or wound current transformers. Window
type was usually used for measuring circuits with less than 600 Volts along with
Bar type; on the other hand the wound type current transformers were used for
measuring over 600 V. The ratio dealt with window type was a little different than
the other transformers since the conductor for this type that carried the magnetic
field was considered the primary winding and there might be additional turns that
might be used according to the demand. Under these circumstances when used
in a circuit to measure current the transformer ratio had to be divided to the turns
to find the actual current value. Bar type current transformer, which was also
used to measure low voltage circuits, was almost the same as window current
transformer but with a bar installed so that voltage readings could be obtained
conveniently as well as current readings. Finally the window type comes with
fixed primary winding and as mentioned before it was used to measure high
voltages.
Consideration of Current transformer sensor as our project’s power measuring
component didn’t seem to be that logical since the other alternatives, both Hall
Effect current sensors and magneto resistive field sensors, appeared to be better
alternatives. Hall effect current sensor provided logic capability, highly linearity,
non-mechanical structure, resistance against heat, perfect to measure high
currents for safety concerns, were insensitive to dust, vibration, humidity, cold
and hot and well-sealed structure. On the other hand magneto resistive field
sensor had good accuracy and high sensitivity. It was durable, it had fast
response rate, had a decent size, could be used in harsh conditions and
reasonable price. Current transformer sensors were very suitable for higher
voltages and their sizes were usually bigger and they don’t provide the
advantages mentioned above.
2.5 Wireless Communication
For communications between devices it was important for the installation to be
feasible. Implementation of a home monitoring system using a wired platform
could not be justified in both cost and time. A wireless technology that could
penetrate walls and communicate securely with a rage of around 100 feet was
necessary. We chose a range of 100 ft due to problems associated with
absorption and reflection of frequency waves. This product was meant to be
used in-doors where there would be several line of sight issues. Choosing a large
range of 100ft should have ensured that the signal loss would not affect the
quality of the product.
33
To explore the options in wireless technologies one must have considered the
characteristics of these devices. For this project the devices most important
attributes were low power consumption and range. This would ensure that the
individual outlets in a home were able to communicate back and forth without the
need for constant maintenance. Other factors such as data rate, speed, and
security were also considered. The wireless technologies researched were the
EZ430-RF2500 wireless development tool from Texas Instruments, the Xbee
802.15.4 RF modules, Arduino Bluetooth technology and the Xbee ZB module.
Overview TI eZ430-RF2500
The EZ430-RF2500 was a development tool that included both the
MSP430F2274 microcontroller and the CC2500 wireless transceiver. This
development tool also comes with additional debugging software and examples.
All of the communication and power monitoring was realized using this
microcontroller in conjunction with the wireless transceiver. By using these tools
together all the microcontrollers used for power consumption readings had the
ability to communicate with each other using 2.4-GHz wireless technology. [9]
Features








CC2500 2.4 GHz, ISM band multi-channel low power transceiver
Supports MSP430 Application UART allowing serial communication to PC
SimplicitTI, low power network stack
Supports development with MSP430 hardware
21 available development pins
Highly integrated, ultra-low-power MSP430 MCU with 16-MHz
performance
Supports development with some 2xx Spy Bi-Wire devices
Supports eZ430-T2012 and eZ430-RF2500T target boards
Figure 21 below of the wireless development tool gave a visual representation of
the hardware. The rightmost part of the picture was the target board; the left part
showed the debugging USB hardware that plugs into a computer. This board
included both a CC2500 2.4GHz wireless transceiver and the MSP430F2274
microcontroller. On the target board pins 3 through 7 were general-purpose
digital I/O pins A0 through A4. General-purpose digital I/O ADC10 pins A12
through A15 were given by pins 8 though 11 respectively. Pins 1 and 12 were
used as ground reference while pin 2 supplies the voltage. [10]
The wireless development tool used SimpliciTI which was a lower power RF
network protocol developed by Texas Instruments. It was generally used in
alarms, controls, meters, or sensors. It utilized up to 8K, 1K RAM depending on
the configuration. It also supported low power features for sleeping devices. This
34
was an important feature in our device as small portable RF networks generally
had onboard battery power. To minimize the power consumption it was
necessary for devices to enter a sleep. Similar to the XBee mesh network the
peer-to-peer network topology had the ability to extend the range of the device
using repeaters. This feature was designed for use in simple small RF networks
with fewer than 100 nodes. Although this technology was not as robust as the
XBee mesh network it was still enough to cover the requirements of our design.
The network protocol supported different communication topologies such as
peer-2-peer and direct p2p through RE. For our design a basic direct peer-2-peer
connection was sufficient. The feature for peer-to-peer messaging could be
accessed using the following syntax.
• smplStatus_t SMPL_Send(lid, *msg, len);
• smplStatus_t SMPL_Receive(lid, *msg, *len);
The MSP430 was especially useful for a project in which sensors were required
for power consumption measurements. To do this we used pinput pins 8 to 11
described above. These pins were important due to their analog to digital(ADC)
conversion characteristics. To measure a sensor using the MSP430 one could
use ADC10 sampling. The ADC10 was a 10 bit ADC which had a total of 2^10 or
1024 values. [11]
The ADC10 was important for sampling analog data from external sensors. It
could convert a voltage or temperature reading into useful information for the
microcontroller. Using this information one could alter the behaviors of other
appliances in the home. Several appliances were monitored simultaneously with
data being shared with a main unit. This unit would receive all the data and sum
up the total current being drawn. This load would be measured against the
current delivered by the generator. If the load drain exceeded the load supply,
then appliances of least priority needed to be interrupted so the generator did not
overload. Overloading the generator may cause appliances in the home to fail.
Enabling the ADC10 was done simply by enabling the ADC10ON bit of register
ADC10CTL0.
The MSP430 allowed for asynchronous communication using UART. By using
UART the microcontroller could communicate with a computer over a serial port.
For the purposes of our project all of the software may be tested with this
wireless development tool. The wireless target boards had an integrated
MSP430F2274 and CC2500 multi-channel RF transceivers. These two parts
could be purchased separately and embedded into their own circuit.
The debugging software included in this package used IAR Embedded
Workbench Integrated Development Environment (IDE) or Code Composer
Essentials (CCE) for writing and debugging software. Both of these packages
included the ability to code in both C and C++. Data was transmitted using
35
UART on the MSP430. UART on this microcontroller was controlled by seven
control registers and one read-only register, each register consisted of 8-bits. [12]
Figure 21: EZ430-RF2500 Development Tool
The CC2500 was a RF transceiver available in several MSP430 microcontrollers.
The advantages of this part lay in its small size and lower power consumption. It
only consumed 13.3mA while in operation. When it went into sleep mode it
consumed as little as 400nA. To exit from the sleep mode it took 240micro
seconds. It could be programmed to transfer data at a range between 1.2 to 500
kBaud. Operating voltages for the CC2500 should have been restricted between
1.8 and 3.6 volts. Due to the transceiver being embedded on the microcontroller
these voltages were supplied by the MSP430. However, the voltages could also
be supplied from a regulated power supply.
Figure 22 showed the input pins of the CC2500 transceiver. As shown the chip
contained 20 pins. Some of these pins were used for ground, it was important to
note that if this transceiver needed to be mounted on a ground plane. This was
the primary chip ground connection. A 26 Mhz crystal was used to drive the
36
synthesizer through the XOSC_Q1 and XOSC_Q2 pins. It also provided a clock
for the ADC. For implementing the CC2500 transceiver on a board a number of
extra components would be needed. These components and their configuration
were also displayed above. The figure above did not include decoupling
capacitors needed to supply voltage to the transceiver. The decoupling
capacitors needed were two 220pF at 10% tolerance rated for 50V on pins 9 and
15 while pins 11 and 14(AVDD) required capacitors of 100nF at 10% tolerance
with a rating of 10V. The values for the remaining components for the CC2500
Transceiver were given in figure 23 below.
Figure 22: CC2500 transceiver and components
Permission requested from Texas Instruments
Communication between a microcontroller and the CC2500 was done through
four input and output pins. The pins used were 20, 2, 1, and 7. Pins 6 and 3
were used to generate interrupts for the microcontroller. [13]
Component
C51
C81
C101
C121
C122
C123
C124
C131
Value
100 nF ±10%, 0402 X5R
27 pF ±5%, 0402 NP0
27 pF ±5%, 0402 NP0
100 pF ±5%, 0402 NP0
1.0 pF ±0.25 pF, 0402 NP0
1.8 pF ±0.25 pF, 0402 NP0
1.5 pF ±0.25 pF, 0402 NP0
100 pF ±5%, 0402 NP
37
Manufacturer
Murata GRM15 series
Murata GRM15 series
Murata GRM15 series
Murata GRM15 series
Murata GRM15 series
Murata GRM15 series
Murata GRM15 series
Murata GRM15 series
C132
L121
L122
L131
R171
XTAL
1.0 pF ±0.25 pF, 0402 NP0
1.2 nH ±0.3 nH, 0402 monolithic
1.2 nH ±0.3 nH, 0402 monolithic
1.2 nH ±0.3 nH, 0402 monolithic
56 kΩ ±1%, 0402
26.0 MHz surface mount crystal
Murata GRM15 series
Murata LQG15HS series
Murata LQG15HS series
Murata LQG15HS series
Koa RK73 series
NDK, AT-41CD2
Figure 23: Component values CC2500
Overview XBee ZB and 802.15.4 RF Modules
The Xbee 802.15.4 RF considered were the regular and pro models. The
advantage of the pro model being that it had a much longer range than the
standard as seen in figure 24. XBee RF modules were interfaced with a
microcontroller through the Data in and Data out ports. These were denoted as
pins 2 and 3 on the Xbee RF module. Using the DOUT and DIN pins serial data
could be sent or received by the microcontroller. The default rate for this
asynchronous communication was 9600 baud or bits per second. To prevent
buffering and framing errors the pins CTS RTS and DTR were used for
handshaking between the Xbee and the microcontroller. This prevented
situations in which the microcontroller could send serial information faster than
the wireless module could process it. The digital input and output ports used 3.3v
of regulated dc voltage.
The disadvantage of the Xbee 802.15.4 modules was in large due to the fact they
did not support extended network range through routing. This was what gave the
Xbee ZB a significant advantage for use in our design. Through using network
through routing the ZB modules could extend the range of a network. On the
other hand the 802.15.4 modules only supported point to point communications.
Similar to the EZ430-RF2500; this wireless solution also included a number of
methods for debugging software through using a PC. This was accomplished by
interfacing the Xbee module to a PC using a USB adapter. This adapter was preassembled and the RF module was simply connected through soldering or by
using a breadboard. The USB adapter would allow communication between the
PC and the Xbee. This device could be powered through a 5 V DC supply. [14]
Xbee
802.15.4
Pro – 802.15.4
RF Data Rate
250 kbps
250 kbps
Indor/Urban Range
100 ft (30 m)
300 ft (100 m)
38
Outdoor/RF Line-of-Sight
Range
300 ft (100 m)
1 Mile
Serial Data Interface
3.3V CMOS UART
3.3V CMOS UART
Frequency Band
2.4 GHz
2.4 GHz
Serial Data Rate
1200 bps – 250 kbps
1200 bps – 250 kbps
ADC Inputs
(6) 10-bit ADC inputs
(6) 10-bit ADC inputs
Digital I/O
8
8
Antenna Options
Chip, Wire Whip,
U.FL, & RPSM
Chip, Wire Whip,
U.FL, & RPSMA
Encryption
128-bit AES
128-bit AES
Figure 23: Component values CC2500
Xbee ZB
The Xbee ZB was used in wireless home applications. It was used for controlling
wireless light switches and to display meter information to displays. The key
features of the Zigbee modules were its low data rate, long battery life, and 128bit AES encryption for data. Another important feature of the zigbee module was
that it could communicate with a total of 65,000 devices. This was more than
enough outlets for any home.
Features:
 Low-cost, low-power mesh
networking
 No configuration needed for outof-the-box RF communications
 Support for larger, more dense
mesh networks
 128-bit AES encryption
 Frequency agility
 Over-the-air firmware updates
(change firmware remotely)
 ZigBee mesh networking
protocol
- Improved data traffic
management
39












- Remote firmware updates
- Self-healing and discovery
for network stability
Low-power sleep modes
TX current:
RX current:
Power-Down current: Industrial
temperature rating (-40C to
85C)
1Mbps Max data rate
2mW output (+3dBm)
(4) 10-bit ADC inputs
(10) digital IO pins
128-bit AES encryption
Local or over-air configuration
AT or API command set
Fully FCC certified
45mA @ 3.3V
40mA @ 3.3V
<1uA @25 degree C
Mesh networking was also supported by the ZigBee protocol. This essentially
increased the range and reliability of our whole network. Instead of each power
strip having to communicate back to a central location, each power strip
microcontroller could relay their information via other power strips. This gave
multiple paths for the data to flow. All of these connections between nodes were
not done through our software. They were built in capabilities of the ZigBee,
allowing for nodes to be dynamically updated and optimized during operation. If
one of the power strips was powered down, then the flow of information would
not be interrupted as it could still be relayed through a different node in the home.
This would allow us to set enable a devices sleep mode without compromising
communication of the devices. Using this technology the range between devices
ceased to be an issue. All homes had several outlets in each room. Using this
mesh network the size of a home would not add any difficulties to the network. In
some cases it could even improve the network as more data paths were
introduced into the system.
UART PIN communication with the XBee ZB as referenced by figure 24 was
done through four of the pins. Pin 3 was used to transfer data from the
microcontroller into the wireless module as an asynchronous serial signal. The
remaining pins were referenced in figure 25. This signal would remain high until
data was transmitted. Data was finally then transferred to another wireless
module through the pin number 2 which was the data output pin. Pins 12(CTS)
and 16(RTS) were used as output and input digital ports. The XBee RF modules
could communicate with any other device that was also compatible with UART.
[15]
40
Figure 24: XBee Module
Mounting
Xbee modules were designed to mount into a receptacle. These receptacles
could be mounted onto a board using through-hole and surface mount
technologies. By using receptacles there was a lesser chance of damaging the
module. All soldering could be done prior to the installation. The modules were
not sensitive to nearby processors or components. Care should have been taken
for installing the antenna in a proper location. Metal objects close to the antenna
would cause reflections that would reduce the efficiency of the antenna. Due to
the nature of this project microcontrollers would be encased in a small enclosure.
An external antenna must have been considered as the on chip antenna would
most likely be insufficient to meet our needs. [16]
PIN
1
2
3
4
5
6
7
8
9
NAME
VCC
DOUT
DIN
DO8
RESET
PWM0/RSSI
PWM1
RESERVED
DTR SLEEP_RQ DI8
PIN
10
11
12
13
14
15
16
17-20
Figure 25: XBee Module Pins
41
NAME
GND
AD4 DIO4
CTS DIO7
ON/SLEEP
VREF
ASSOC AD5 DIO5
RTS AD6 IO6
AD3-AD0 DIO3-DIO0
Arduino Bluetooth Technology
Features:
 Microcontroller
 Operating Voltage
 Input Voltage
 Digital I/O Pins
 Analog Input Pins
 DC Current per I/O Pin
 DC Current for 3.3V Pin

Flash Memory



SRAM
EEPROM
Clock Speed
ATmega168
5V
1.2-5.5 V
14 (of which 6 provide PWM output)
6
40 mA
50 mA
16 KB (of which 2 KB used by
bootloader)
1 KB
512 bytes
16 MHz
The Arduino Bluetooth was a wireless technology using the WT11 class 1
Bluetooth 2.1 module mounted alongside an ATmega168 microcontroller. The
Bluetooth modem could also be purchased separately and then connected to an
Arduino board shown in figure 26. This was not an issue related to our project
since all of our components would be connected using a professional PCB
layout. The board components could be powered with a voltage range between
1.2V and 5.5V. This pin should be supplied with a voltage of around 3V. Voltages
outside of this range would damage the board. The voltage was supplied to the
board via the 9V input pin. The 9V pin cannot exceed 5.5V. This voltage could be
supplied from the V+ on board DC-DC converter or it could also come from a
regulated 5 volt power supply.
The device comes with 14 digital pins for both input and output. These pins
could be accessed using functions defined as pinMode(), digitalWrite(), and
digitalRead(). No pin could receive more than 40mA of current. These would be
the pins used in interfacing the Arduino board to the Blueooth modem.
Developing the wireless interface for the Bluetooth modem required a Bluetooth
dongle for the computer and a Blueooth modem for the Arduino board. These
two devices could be used in the debugging phase to communicate with each
other. As with the TI eZ430-RF2500, Bluetooth technology could also be used as
part of a development tool for the wireless software. Once the software was
tested and fully functional then each board would be loaded with the software so
devices could communicate with each other autonomously. The final design
would not require a computer.
42
Communication for the Bluetooth Arduino device was accomplished via the Serial
port. Serial port 0 (RX) was used to receive TTL serial data while Serial 1 (TX)
was used to transmit data to other devices. In addition to the RX and TX inputs
the Bluetooth modem needs an additional PWR and GND pins to be connected.
To initiate wireless communication we would need to access the serial
communications port using the code.
void setup() {
pinMode(ledpin, OUTPUT);
Serial.begin(9600);
}
Implementing Bluetooth communications using the Arduino board was extremely
easy given the amount of information that was shared. Range with the Bluetooth
WT11 could be an issue. Shown below the size of the Arduino was also an issue.
Comparing this picture to the development tool of the TI eZ430-RF2500 one
could see that there was a significant size difference between the device
components. To take this design to the next step all of the components used for
the power metering and control applications would need to fit inside a
conventional power outlet.
43
Figure 26: Arduino Uno
For this project all of the discussed wireless technologies were possible. The
Xbee ZB platform was widely used in home power metering applications. The
implementation of this method may not be as challenging. In addition the node
jumping characteristics of the Xbee ZB increase the range and versatility of that
platform. Texas Instruments solutions to home metering applications were harder
to implement but they were also cheaper.
The easiest to implement of all the technologies researched was the Bluetooth
modem using the Arduino board. Since the Arduino board was an avid
development platform among hobbyist there were many examples showing how
to interface the two components. The biggest downside to this option was the
cost. Arduino boards cost around $30 dollars. For the testing phase part of our
project we would need a total of 5 boards. There was an advantage of not using
the Arduino Blutooth. $115 could be saved when you compare the cost of the
boards to the cost of parts we could sample for free from Texas Instruments.
Our final design choice for the wireless development of our project would be
using the MSP430 with the CC2500 transceiver. This was due to its embedded
nature of the C2500 transceiver and low power consumption. The cost and size
of the MSP430 were also important features we considered. Given the amount
of room for a conventional power outlet the production design needs to be small
enough to fit inside of it. The implementation of these two components would be
much more difficult than the technologies discussed earlier. Due to the lack of
information regarding these components we feel there was a greater opportunity
to learn. [17]
2.5.6 Antenna Options
All of the distance specifications shown on each part were given by using
controlled experiments. It does not take into account the loss of signal due to
absorption, reflection, and scattering of the waves. Metal surfaces could reflect
signals. Other enclosures around the wireless module would absorb frequency
signal. Due to these natural factors different antenna styles should be
considered.
There were many antenna options for each of our wireless solutions. These
antenna options give us the ability to scale our design. The considered wireless
transceivers come with an on-chip antenna which was unlikely to meet our
project specifications. While these antennas offer high portability they would not
provide a strong enough signal for the design. Given the design specifications
the transceivers would be contained within an enclosure. This enclosure would
reduce the effectiveness of the on-chip antenna. A larger antenna inside the unit
may be needed.
44
Wire whip antennas similar to those seen on older cars offer a longer range of
communication but must be visible for high performance. This antenna option
had a clear advantage over the on-chip antenna due to its longer range. However
it does not provide much portability. Aesthetics would also be an issue. A
consumer would prefer a hidden antenna over an unattractive whip antenna.
Given the time constraints the project would be developed allowing form to follow
function. Other antennas such as those used in laptops were more attractive but
harder to implement.
Wireless laptop technology uses a U.FL, RPSMA antenna. This solution offers a
small footprint for PCB mounting. When attached to this Male connection a
female coaxial RF antenna could be used to boost Wi-Fi signal. This antenna
does not have to be exposed outside of the unit.
For the initial prototype of this project aesthetics would not be an issue if the onchip antenna was not powerful enough to meet the design specifications we
would mount a whip antenna outside the unit to meet the range required.
2.6 Microcontrollers
Each block of our design would incorporate a microcontroller, and each of these
would communicate with a central, “hub”, microcontroller which would be
physically connected to the LCD screen output. Each microcontroller would need
to take in data from its source, analyze that information, and wirelessly transmit
the necessary data to the hub microcontroller. The microcontroller associated
with the Automatic Transfer Switch (ATS) module would need to communicate
which power source was currently being used for the house. The microcontroller
connected to the Generator module would take the information from the
current/voltage sensor and tell the hub the maximum supply voltage available
from the generator. The microcontroller connected to the Power Strip module
would analyze the power measurement data from each of the three outlets on
our power strip, and submit this information to the hub microprocessor. This
microcontroller would also need the ability to turn on or off any and all of the
outlets, based on the supply needs of the user and the power supply capabilities
of the generator (when in use). Finally, we would need a centralized
microprocessor at the LCD module which would take in all the pertinent data
from all other modules as well as inputs from the LCD user display, and transmit
orders to the power strip microcontroller accordingly. In addition, it would need to
relay information back to the LCD user display. In total, this give four separate
microcontrollers, each with its own wireless transceiver, capable of
communicating to and from the User Display module. These concepts as they
45
relate to the overall project, in regards to the individual microprocessors, were
portrayed below in Figure 27. Ideally, all of these microcontrollers should use
minimal power, and have the ability to remain in a standby mode for long periods
of time, as well as wake up quickly.
Choosing a microcontroller among the many options available to us was a
difficult task, as there were many viable options. As the technology expands and
develops, more complex microcontrollers that have many more functions were
available for minimal costs. We thus wanted to simplify and break down our
specific needs as concise a form as possible, to make the best choice clear. Our
microcontrollers must be low power consuming, compatible with our chosen
wireless interface as well as each module in our main block diagram, and with a
fast response and relay time to minimize the time that our end user spends
without power to his home.
Figure 27: Microcontroller block diagram
46
2.6.1 MSP 430
One of the microcontrollers considered was TI’s MSP430 family of 16-bit RISC
mixed-signal processors, shown below in Figure 1. In addition to the great
support from TI for the engineering students at the University of Central Florida,
the MSP430 offered tremendous power and a wide array of applications and
capabilities for very low power consumption. The MSP430 offered many built-in
features for the monitoring and metering of all utilities, including one-to-three
phase electricity metering. In addition, TI boasted having the “world’s lowest
power microcontrollers” in its MSP430 line, essentially a high performance-topower ratio. A low power demand from the microprocessor was ideal for our
project, which would allow the generator to support as many electronic devices
as possible when in use. The MSP430 also contained several built-in low power
modes, where the processor drained virtually no power, and could be woken up
with just a 1 micro-second delay. These modes varied slightly from each other,
and could be easily chosen between based on the needs of each of the
microcontrollers in our project. As shown below in Figure 28, the MSP430
microprocessor available today was an improvement from previous versions and
other products for single phase metering applications, due to the consolidation of
many parts into one, and also very easily connected and communicated with an
LCD display, as would occur in our project. This would work well for our group,
as dual and three phase metering was not necessary in a residential home.
Figure 28: Microcontroller
Permission requested from www.ti.com
The seven (7) LPMs (Low Power Modes) of the MSP430
optimization of coding. With the sub-1 µs wakeup time
interrupt sources in combination with the LPMs, only the
and peripherals that were needed would be used. The
modes were as follows:
47
allowed for increased
and several different
corresponding clocks
low power operating





Ultra-Low Power Active Mode: Down to 120 µA/MHz @ 2.2V
Standby Mode with self-wakeup & RAM retention (LPM3): Down to 0.7 µA
@ 2.2V
Standby Mode with self-wakeup (LPM4): Down to below 100nA @ 2.2V
Shutdown Mode with RAM retention (LPM3.5): Down to below 100nA @
2.2V
Additionally, the MSP430 had instant (less than 1 µs) wakeup time from
Low Power Modes, as well as an always-on zero power brown-out rest.
Texas Instruments had available a line of MSP430 microcontrollers built
specifically for the purposes of utility metering. This series, the MSP430AFE2xx
series, includes low-power 16-bit analog-front end (AFE, which contains a tuner
and an analog-to-digital converter, and which here was specifically used for
measuring) MCUs.
2.6.2 Arduino Uno
A second microcontroller brand that was considered was the Arduino line of
microprocessors. This was an open-source platform, which was designed to be
easy to understand and provided both amateur and professional engineers a
common platform to design and create projects to their needs. Due to its opensource nature, support for this microcontroller could be hit or miss. Typically,
when you ran into an issue, someone else had run into that same issue before,
and it was only a matter of time spent scouring the forums for the correct answer
to suit your needs. However, this could be a very difficult, tedious process, and it
may also turn out that the answer did not exist. Thus, reliable support would be a
concern with this choice. The Arduino also tended to be a cheaper alternative
than other microcontroller platforms available, which would definitely be a benefit
for our project. Additionally, the programming and both the software and
hardware that would be developed for this board were open source, allowing for
optimal personalization and cost-saving measures. The software ran across all
major platforms (Windows, Macintosh OSX, as well as Linux), which was a good
feature to have to allow work at nearly any computer. However, our group had
access to Windows computers on a regular basis, so this benefit was limited. [17]
Our group considered microcontrollers in several I/O board layouts from the
Arduino line. First, the Arduino Uno was the standard layout, with a USB port, 14
I/O pins, and 32KB of flash memory. This board could be powered through the
USB connection, and the power source was selected automatically. Additionally,
the board could be powered by an external battery supply, either into an AC
adapter or directly into a battery. Its operational voltage was between 6 and 20 V,
though it was recommended to operate at >7V; below this value, the board may
become unstable. Conversely, operating past 12V may cause the board to
overheat. Therefore the recommended range was between 7 and 12 V. This
information was summarized below, along with other pertinent data, in Figure 29.
48
Factor
Data
Microcontroller
Operating Voltage
Recommended Vin
Limit of Vin
Digital I/O Pins
Analog Input Pins
Flash Memory
Clock Speed
ATmega328
5V
7V-12V
6V-20V
14
6
32 KB (31.5 available)
16 MHz
Figure 29: Arduino Uno Specifications
Permission requested from www.microchip.com
There were four power pins: VIN, 5V, 3V3, and GND. The 5V pin was used to
power the components on the board, and the 3V3 pin provides a 3.3 V power
supply. The Arduino Uno had 14 digital I/O pins which operate at 5 volts, and
able to receive a maximum of 40 mA current, and each with an initially
disconnected internal pull-up resistor (valued between 20 and 50 kΩ.
A more powerful board than the Arduino Uno, the ATMega 2560 was a board
layout and processor that would be one step up in terms of raw computing
power. This board had 54 digital I/O pins, 16 analog inputs, 4 hardware serial
ports (UARTs), and a 16 MHz crystal oscillator. The board had 256 KB of flash
memory, which was a significant increase from the Uno (32 KB). More memory
would allow our group to accomplish more tasks as well as store more
information for processing before having to clear space. This could allow for
more flexibility in the programming of each of the MCUs, as well as in the data
retrieval by the main processor in the LCD. Additionally, the Mega 2560 (which
was a 2nd generation controller in this line) could be powered either through the
USB port or an external power supply. This extra feature may come in handy in
allowing more flexibility but would more than likely not be used, as this
microprocessor would be powered independently of our LCD (which may use a
USB connection).
2.6.3 Microchip PIC
A third microcontroller considered was the PIC (short for Peripheral Interface
Controller) manufactured by Microchip. The PIC microcontroller was available in
a variety of configurations and specifications in 8-, 16-, and 32-bit sizes. There
were several draws for using the PIC microcontroller. First, there were 9
programmable sleep modes, during which the PIC draws less than 1 µA. Five of
these modes were a “deep sleep”, which was a preferable choice for
programming applications with long periods of inactivity, and which includes 7
possible wake-up sources. These sources included: Brown-out Reset, Real-time
Clock, Reset, Timer, Interrupts, and Power-on Reset. Additionally, in direct
49
comparison with the MSP430 series of MCUs from TI, the technology in the PIC
microcontroller required 13% less current in its sleep mode for a real time clock,
and between 33% and 178% less current in its deep sleep modes. Only in the
Time mode was the current less in the MSP430. This information was shown
below in Figure 30.
Figure 30: MSP430 vs. XLP
Permission requested from www.microchip.com
Second, the PIC MCU ran on very low power, which would be useful for our
project. Sleep currents for the PIC run as low as 20 nA, and active mode currents
could be as low as 50 µA/MHz. Also, more than 80% of the basic instructions
could execute in only a single clock cycle. This could help to reduce the time
needed when multi-tasking; for our project this means that nearly as soon as the
user specifies which utilities and outlets were to be turned on (when running on
generator power), this information could be relayed to the other microprocessor
and, within 2 clock cycles, the outlets could be active. [7]
The mid-range architecture PIC which our group would be interested in had
between 8 and 64 pins, a single interrupt ability (which may limit the capabilities
of our project), up to 368 bytes of data memory, 14 KB of program memory. The
MCP3905 was an energy metering integrated circuit available from Microchip
and was designed specifically for use with the Microchip PIC. It was capable of
supplying active power measurement for single-phase, residential energy
50
metering, and would thus be very useful for our project. The IC features an easyto-use capability to supply a frequency which was directly proportional to the Pave,
or average active real power, as well as less than 0.2% error on most
applications.
As none of the group’s members were sufficiently fluent in other programming
languages, another concern for choosing a microcontroller was the capability to
program the MCU in C, C++, or (more likely) a combination of the two. Having a
microcontroller with this capability would greatly reduce the amount of time
required for the group’s members to learn to program the necessary functions
and code for the signal processing and data manipulation portion of our project.
The MSP430 uses a combination C and C++ languages in all its programming,
and would thus be a valid choice. The Arduino uses its own programming
language, whose syntax was very similar to C, and thus would pose only a small
challenge to learn. Finally, the Microchip PIC used an assembly programming
language, similar to that used in the MC68HC11 microprocessor. However, there
was a programming package available from the manufacturer (Microchip) that
allows design-users to program in a friendlier, C environment.
2.7 PCB Fabrication
The final printed circuit board needed to be durable enough for our presentation.
There were many hobbyist options when it comes to creating your own PCB
boards. We would consider some of these options as other methods were not as
cost effective and take longer to process. After concluding the test procedure a
whole mockup of the design would be implemented using one of the following
design options.
The first option included creation of the PCB using a permanent waterproof
marker and an acidic FeCl3 solution. The advantage of this option was that it
used materials which could be found locally. Implementation of this design could
take hours as opposed to the days or weeks it could take to receive a
manufactured PCB. Using this method a layout could be created in Photoshop
or Microsoft Paint. After your circuit layout was printed you simply place it on top
of a copper sheet and mark where the holes should be. Now remove the
template and simply trace the connections on the copper sheet with your marker.
Applying a coat of ink right before the copper board goes into the FeCl3 solution
adds a layer of permeability. The solution would not eat through the ink; instead
it would corrode all of the copper around your connections. Wash off the PCB
thoroughly and then mount your parts onto the board by using a solder. Small
parts should be mounted first. [18]
The problem with this hobbyist option was that drilling the holes for such a small
component as an MSP430 would not be possible. The pin connections were too
close and the risk of damaging the board was significant. This method seemed
to be most effective when you were building a circuit which does not implement
51
components that were small and fragile. This option could be used in our design
to connect the larger components. Smaller circuits for components such as the
transceiver and the MSP430 would be created using professionally manufactured
boards. These companies accept designs through their website and have a
turnaround time of around one to two weeks.
Online vendors such as pcb.com or expresspcb.com offered a professional PCB
product. By using a third party vendor we could be certain that all the
components were mounted correctly and that there were no erroneous
connections made between the components. These companies offer high quality
fabrication at a higher cost. Although processing time and cost were higher we
were benefitted with some advantages in reliability and accuracy. 4pcb.com
currently offers a free PCB with the order of four or more PCB at a price of $33
each. This area would be enough to meet the design for all of our circuit
components and different units. By printing the same design on each board we
could save in fabrication costs. In total we would print five boards which could
create 10 circuits. Three in total for each power strip along with one for the main
circuit and an additional current sensing circuit near the generator. Extra boards
would be included in the order in case we run into any problems.
Due to time constraints we would order an initial test board as soon as a design
was finalized. This would allow us to test the circuit for a single power strip. If
the circuit passes the testing phase we would order other identical circuits to
build the remaining circuits.
To design these circuits companies required designs which were created using
specific file formats. In the case of pcb.com they accept gerber, drill, and drill tool
list files. Other formats such as ODB++ and AutoCAD were also accepted. There
were several programs that offer the ability to create designs in these file
formats. The different options for creating the PCB layouts were explored below.
[19]
2.8 PCB Software
To create our PCB we would need a schematic diagram. We would consider two
different options for PCB software. The first option for our PCB design would be
using EagleCAD.
It contained a large library of parts from various
manufacturers. Limitations with EagleCAD include size and layers. These
limitations were lifted if you purchase the professional version of the software. At
around $1500 the processional version was out of our budget range. The light
version of the software was only $50 but it imposes limitations on the amount of
schematic sheets, layers, and the area size. The area allowed using the light
version of the software was only 100x80mm. These restrictions should not be a
problem because our design was small enough to fit inside the size
requirements. In addition the two layers provided would be enough. In choosing
between the two software packages the only decision we need to make was in
52
the cost and feasibility. The most important feature with the paid version of
EagleCAD was the parts ordering process. After the design was finished this
feature allows us to order all of the parts in the project at once. This feature was
very attractive since we cannot afford any delay during this part of the project.
The second option would be using FreePCB; an open-source PCB editor. As
seen below in figure 31 the FreePCB software was intuitive and very easy to use.
It provided up to 16 copper layers and we could design a board of up to 60 x 60
inches. Users have the ability to add pins traces and stub traces visually. The
menu on the left gave quick access to select different masks. By selecting
different masks users could edit pins or traces without altering other components
of the PCB. It turns out that this software was very similar to EagleCAD but with
the limitations removed. The downside to the FreePCB software was that it
comes with a limited library of parts. It also lacks the ability for a user to purchase
all of the parts needed for the design through the software. [20]
Upon considering the limitations of each PCB software we felt confident that the
better choice would be EagleCAD. Due to the requirements of our microcontroller
design the hobbyist version of EagleCAD costing $50 would be enough to
develop the design. This version of the software imposes heavy limitations on the
standard features. These limitations would not cause any issues with our design
since the design could be developed within these constrictions.
Figure 31: Free PCB Software
53
2.9 Automatic Transfer Switch
An automatic transfer switch was a device commonly used to switch an electrical
loads power source from normal to emergency power during a time of normal
power failure. This ensured a minimal disruption in available power and allows
the consumer to keep on utilizing whatever electrical appliance they were using
before power loss. As soon as normal utility power was restored the automatic
transfer switch senses this change an then switches the load back to normal
power.
Our project would demonstrate the concept of using our power strip module to
automate and monitor an entire house. However due to cost, time, and location
this was impossible so we were demonstrating the overall concept using a smallscale model. To replicate normal and emergency power we would utilize the
common receptacle outlet seen in any classroom. These receptacles were rated
for 120 volts, 20 amps and supply more than enough power for the purpose of
demonstration. With the availability of 20 amps coming from the receptacle we
had to ensure the transfer switch we selected was rated for it. This was why we
chose to use the Bodine GT20A transfer switch in our project. This transfer
switch seen below in figure 32 was normally used for emergency lighting in
commercial buildings, but it was inexpensive and was perfect for the
requirements of this project.
Figure 32: Automatic Transfer Switch
54
The overall design of the transfer switch was out of the scope of this project. We
were simply adding it in for demonstrating purposes so the audience could
visualize the switch between normal and emergency power and the effect it had
on the performance of the system. Shown below was figure 33 and it was the
typical wiring diagram for normal transfer switch operation.
Figure 33: ATS wiring schematic
Permission requested from Bodine.com
The block that says “Emergency Ckt Brk Panel” would be our emergency power
that we bring to the transfer switch from the wall receptacle. A normal extension
cord with (3) 12 gauge wires would be used which meet NEC standards. The
block that says “Normal Ckt Brk Panel” would be our normal power that also
comes from a receptacle using the same type of extension cord. The last block
that says “Lighting Load” would serve as the feed to our “house panel”. If this
were an actual scale model this would be the connection to your houses main
electrical panel. We would use a multiple outlet power strip to serve as our
“house panel” and our power strip modules would plug directly into the multiple
power outlet strip. This would give the user the overall concept of the home
power system using a small-scale model.
Referring back to Figure --- you could three terminal located on the right side of
the automatic transfer switch and they were labeled 5, 6, and 7. These terminals
were used for low voltage controlling of the automatic transfer switch. For
example if this transfer switch were used in a commercial application where low
voltage switching was utilized the signal from the switch would connect to
terminals 5,6, and 7 on the transfer switch. The actual wiring configuration
depends on what you were trying to achieve with the switching. We would be
utilizing these terminals to connect a wireless transceiver to them. After reading
through the user’s manual of the transfer switch we found that during normal
power operation terminals 5 and 6 would measure 12 volts between the two.
55
When normal power was lost and the transfer switch switches over to emergency
power the 12 volts between terminals 5 and 6 was lost. This feature would serve
as the interrupt to our touch screen controller to inform the program whether it
was operating in normal or emergency conditions.
2.10 Optical isolator
An optical isolator was basically a transfer device. It was generally used in
electronic circuits to secure the more expensive equipment like micro controllers
in the expense of its own life. This usually was not a big deal since an optical
isolator was a cheap component. The protection it provided was against voltage
and current spikes which would harm expensive equipment. As seen in the
schematic representation below it was formed of two parts with a gap in between.
In the first part there was a light emitting diode and in the second part there was
a photo sensor separated from each other with a gap. Basically this gap was
what secured the expensive equipment from the surges since there was no
conducting element between the LED and the photo sensing diode. The way it
functions was once there was electrical input signal LED lights up. This light can’t
be seen from outside because it happened inside the component in a dark
environment. This light was received by the photo sensor and was converted to a
proportional electrical energy. If there were a voltage spike only LED would be
harmed and the other side of the optical isolator would remain unharmed. Usually
opto-isolators shown in figure 34 could withstand 10kV though without breaking
down and secure the circuit. Another characteristic property of optical isolator
would be its being a half way rectifier.
Figure 34: Schematic of an Optical Isolator
Public domain no permission needed.
While designing the voltage sensing circuit, it was the first time us as a group got
introduced to this component. In every forum we read about wiring
microcontroller in an electronics circuit experienced electronics designer
engineers were advising the usage of an optical isolator to protect the heart of
the circuit; the micro-controller. After the voltage divider we decided that it would
be wise to put an optical isolator since this circuit was not limited to the Hall
voltage as it was in the current measuring circuit but it was connected to the main
grid where it was vulnerable to rapid voltage spikes.
56
Section 3: Design
3.1 Touch Screen Monitor
After careful considerations of the needs and goals of this project we found that
the QScreen Controller tm from Mosaic Industries, Inc. best suited those
requirements.
The QScreen Controller tm was a combination of a Cprogrammable single-board computer with a touch screen operated graphical
user interface. This touch screen monitor was powered by the Motorola 68hc11
microcontroller, which we were very familiar with from the laboratory section of
Embedded Systems class. The fact the each of the group members was familiar
with the C-programming language as well as the assembly language of the
Motorola 68hc11 was what lead us to selecting this product for our project. [1]
Because this project was a proto-type of product that would be utilized in
consumer household, appearance and size were of great importance and
concern. We envisioned this touch screen controller to be mounted near the
houses existing thermostat so for cosmetic purposes we want our display to be
about the same size as a standard thermostat. After taking measurements of a
few thermostats we found the average to be 4” tall and 6” wide. The QScreen
controllers overall dimensions were 4.125” tall and 6” wide making it the ideal
candidate for size requirements. A picture of the overall appearance and design
of the QScreen controller could be seen below in Figure 35. As you could tell
from the figure below the QScreen controller would surface mount perfect in a
household wall with a cutout not much bigger then needed for a typical switch or
receptacle outlet box.
Figure 35: QScreen Controller
Permission requested from Moasic Industries
57
The QScreen controller was equipped with a touchscreen operated graphical
user interface on a high-contrast 128X240 pixel display with a 5X4 touchscreen
overlay. The screen itself measures 4.8” diagonally and was a high contrast
CCFL white-on-blue monochrome LCD. The user did have the ability to control
backlight and contrast of the screen using software provided with the product.
Because of the ability to navigate through different menus this size screen was
more then enough to satisfy the requirements of this project.
As noted earlier the main reason for using the QScreen controller was having
previous knowledge of the Motorola 68hc11 as well as the C programming
language.
The QScreen controller gives the programmer the option of
programming in either Control C or QED-Forth. We chose to use Control C as
our programming language for the programming of this project. Control C tm
cross-compiler comes with the purchase of the QScreen controller was written by
Fabius Software Systems and customized by Mosaic Industries to facilitate
programming the QScreen controller in C. It was a full ANSI C compiler and
macro pre-processor. The compiler supports floating-point math, structures and
unions, and allows you to program familiar C syntax. Programming was easy
done using an interactive debugger and multitasking executive. The controller
was pre-loaded with hardware control routines including drawing functions for the
display. After programming in ANSI C by compiling our code on our PC and
downloading the code to the QScreen Controller it would then be automatically
executed. The real time operating system and onboard flash memory manages
all required initializations and auto-starts our application code.
The Qscreen controller was not only our graphical interface to the home user, but
it must also have been able to receive signals from multiple power strip modules
through wireless technologies. The controller must then take the received data
and process the information so it could be accurately displayed on the screen to
the home user. This task was easy done using the dozens of real time analog
and digital I/O lines on the QScreen’s Motorola 68hc11 microcontroller. The
68hc11 controls eight 8-bit A/D lines, 8 digital I/O lines including timer-controlled
and PWM channels, and two RS232/485 ports. Two of these analog inputs and
outputs would be utilized for the wireless RF transmission of the data being sent
and received. It was important to verify compatible voltages on the input and
output pins of each device in figure 36 show below.
58
Figure 36: CPU I/O
Permission requested from Moasic Industries
One of the most impressive features of the Qscreen was the pre-coded I/O
drivers provided for all I/O, and this makes it easy to do data acquisition, pulse
width modulation, motor control, frequency measurement, data analysis, analog
control, PID control, and communications. Most likely we would utilize some of
the analog inputs to receive wireless transmissions from a TI C2500
transceiver/receiver. The processor was more than powerful enough for the
application of our project and was a good selection because the project could
easily have options added on in the future.
3.1.1 Display Layout and Design
Now that we had selected the touch screen monitor for the project we need to
design the functionality and appear of the interface to the user. We wanted to
avoid a busy screen and keep it simple and easy for the user to read. Multiple
pages would be avoided so as not to lose the user in navigation throughout the
program. A preliminary touch screen output could be seen below in figure 37.
As you could see the legend indicated that a boxed area with the dotted
background would be the only input area for the user. These areas included
adding a module, turning modules on and off, and also the ability to input the
power rating of their home generator.
59
ADD MODULE
MODULE #
GENERATOR
KW RATING
ON/OFF
CURRENT
KW
KW
AVERAGE
KW
DAILY
COST
MONTHLY
COST
1
2
CURRENT GENERATOR CONSUMPTION
KW
REMAINING GENERATOR POWER
KW
Figure 37: LCD
We want to make sure the product was expandable to the user, which was why
we included the “Add Module” feature. This allows the user to make the initial
investment of the touch screen controller and maybe one or two power strip
modules to test out and use in their house. At any time in the future they could
always purchase another power strip module and connect it to their existing
network. All the user needed to do was plug in the power strip module wherever
they would like to use it and then simply walk to the touchscreen monitor and
click “Add Module”. This would automatically recognize the power strip module
just plugged in and would assign to it the next number in the module list. It was
important to note that if the user wished to add multiple power strip modules at
the same time that they would have to plug each one in separately and click the
“Add Module” button before plugging in the next. The reason for this was when
the user clicked “Add Module” the program was going to be looking for new
devices and if more than one was available an error might occur.
The next feature the user would be able to control was turning the power strip
modules on and off. This option could be used during normal power operations,
but was mainly intended to be used during times of emergency generator power.
During normal power operation the user should have all the modules powered on
and simply turn off their appliance how they normally do. The only advantage to
using the on and off feature during this time was to conserve power on appliance
that draw power even when they were off such as televisions, computer, etc.
This way you could shut the power off to them via the touch screen without
having to unplug each power cord from the wall. The main purpose of the on and
off feature was used when the system was in emergency mode. As soon as
normal power was lost and the transfer switch switches over to emergency mode
60
the QScreen controller would automatically shut off all power strip modules and
wait for the user to decide what modules they want on. By looking at the
“average power consumption” column the user could identify normal
instantaneous power draw from individual appliances measured in KW. The
QScreen knows the KW rating of the generator since the user initially input its
value. As the user turns on individual power strip modules the program would
calculate the remaining available generator power and report it back to the
screen. This way the user would be able realize when they were approaching
the limitations of the generator. Safety features would be added to insure the
user does not accidentally turn on an appliance that would overload the
generator causing failure.
The user would also have the ability to enter the KW rating of their home
generator. This was important for the “emergency mode” operation of project.
The software would monitor the usage of power from the generator as well as all
the power strip modules and would compare to make sure the power
consumption of the all the power strip modules together does not exceed the
rating of the generator. If this were to happen the generator would fail and the
project would have not satisfied its goal.
The Qscreen was pre-loaded with GUI development software to help aid in the
design of the Qscreen output. Most programmers began with sketches of the
screen and eventually worked their way into programming once the design look
was achieved. The LCD device drivers provided 2 methods for writing graphics
objects to the screen. Graphics may either be placed in the graphics array stored
in RAM or sent directly to the LCD’s RAM. The designer had the ability to create
their own symbols and graphics, but they may have also used the generic
symbols pre-loaded with the GUI software. Figure 38 below showed these
available pre-loaded symbols.
Figure 38: Preloaded controller graphics
Permission Requested from Moasic Industries
61
When graphics data was placed in the graphics array the Update_Graphics
function must be called to make the modified graphics array appear on the LCD.
This technique was faster than drawing many different graphics and buttons
individually to the screen directly. Menus and static graphics were displayed this
way. This was the standard method of placing graphics on the screen.
Direct drawing to the LCD was useful when you only want tochange a small
portion of the screen. When drawing only a small number of objects covering a
small part of the screen, direct drawing was faster than using the graphics array
and the Update_Graphics function since only the affected portion was updated.
Button presses use direct drawing since the buttons were single objects and
occupy a small part of the screen. This provides a much faster response than
having to draw to the graphics array and then update the entire display. Figure
39 shown below was supplied with the purchase of the Qscreen and was used as
in aid in the design of the GUI interface.
Figure 39: LCD geometry
Permission requested from Moasic Industries
62
3.1.2 Programming
The Qscreen controller could be programmed in C or FORTRAN programming
languages. All of our group members were familiar with the C programming
language so we would utilize it in the software design of this project.
To begin the first piece of software that needs to be written deals with
recognizing power strip modules throughout the home network. When a user
plugs in a module and goes back to the touchscreen to hit the “Add Module”
function the software would poll the network to search for new devices. This
could be achieved by programming the wireless transceiver in the microcontroller
to send out a pre-defined signal indicating it was new to the network. Upon
finding the new module the software in the touch screen controller would assign it
an address and then proceed with normal operations. To realize this feature a
more in-depth study of the communications protocol of the TI C2500 must be
done as well as the A/D conversions in the main microcontroller. These studies
were currently underway and involve a cumbersome trial and error approach, as
none of us were computer engineering or computer science majors. We were
confident however that we would be able to implement the programming in this
project over the next few months.
After power strip modules had been added to home network the QScreen
controller was now ready to take samples from all of them and analyze, store,
and display the result to the touch screen. In order to accomplish this we were
going to have to receive packets of information in sequence from each of the
power strip modules. At this time we were unsure of the bit size of each of the
packets sent individually, but regardless of the size of the packet the concept
remained the same. The QScreen would mostly be receiving information from the
power strip modules, but it also must be able to send a signal to either turn on or
off the relay controlling the receptacles in each individual power string monitor.
To achieve this every time information was received from a power strip module
the program would determine whether or not the on/off button was pressed. If the
button was pressed the program would tell the controller to send a signal to the
module turning on or off the relay and then would proceed to receiving
information from the next power strip module. If the on/off button wasn’t pressed
the program would simply move on to receiving information from the next power
strip module. Below in figure 40 was a flow chart showing how the software
would execute receiving/transmitting information from all of the power string
modules.
63
QSCREEN
RECEIVE
MEASUREMEN
T
R
STORE DATA
IN MEMORY
CALCULATION
S/UPDATE
SCREEN
YES
EXECUTE
ROUTINE
CHECK
TOUCHSCREEN
FOR
INTERUPTS
NO
MOVE TO
NEXT
MODULE
Figure 40: Program flow chart
It was very important that the design of the touchscreen controller includes a
battery backup because all of the information from each individual power strip
was store in its memory. A sudden lose in power could potentially wipe an entire
month’s worth of data.
3.2 Power Measuring and Circuits
Measuring the electrical power of a given circuit was the most critical and
fundamental concept of this project. Not only were we concerned with allowing
the user to view an accurate power consumption and cost analysis, but we were
also using power consumption analysis to determine the safe operating loads of
a home backup generator.
There were many different methods to measure the electrical power of a given
circuit. For instance the Hall Effect sensor used a current perpendicular to a
magnetic field to measure voltage, current, and power levels. Another way was
to use voltage response measurements where the electrical characteristics of a
circuit were determined from the amplitude and phase of a test current flowing
64
through the circuit. To measure power as stated previously we needed a sensor
to measure the phase difference between the voltage and the current. It was
important for safety purposes for us to consider the voltage and current
magnitudes of the circuits to be measured. As shown below in figure 41 a very
small amount of current could be fatal if short-circuited across the body. This
problem demonstrates the advantages of using the Hall Effect sensor where the
output voltage of the sensor was much smaller in proportion to the input signal
therefore protecting the user from a potentially deadly shock.
Figure 41: Shock hazard safety chart
Permission requested from www.osha.gov
Our power strip modules as well as the module that monitors the generators
power output would utilize the same power measuring circuit. For us to measure
and calculate the output power of a given device we need to measure and record
the voltage and current waveforms of that device. The voltage and current
waveforms coming from the utility grid were analog signals, but our
microcontroller used for data management and calculations only deals with digital
signals. To overcome this problem would be using analog to digital convertors.
Our application requires one measurement device to capture the voltage across
the terminals of a load, and the second device to capture the current going
through the load. However, the actual power calculation depends on the resistive
and reactive components in the circuit. In order to have an accurate power
measurement, we find it necessary to study the electric power a little bit more
and define its different components. In an alternating current power system,
energy storage elements such as inductance and capacitance introduce periodic
reversals of the direction of energy flow. The portion of power flow that averaged
over a complete cycle of the AC waveform, results in net transfer of energy in
one direction was known as real power. On the other hand, the portion of power
flow due to stored energy, which returns to the source in each cycle, was known
as reactive power. The utility company does not charge residential customers for
65
the reactive power drawn from the circuit because this energy was conserved
and was returned to the source. As shown below in figure 42 alternating current
power systems consist of three types of power.
NAME
UNITS
EQUATION
Real Power (P)
Watts (W)
𝑃 = 𝑉 ∗ 𝐼 ∗ 𝑐𝑜𝑠𝜃
Reactive Power (Q)
Volt-Amp-Reactance
(VAR)
Volt-Amp (VA)
𝑄 = 𝑉 ∗ 𝐼 ∗ 𝑠𝑖𝑛𝜃
Apparent Power (S)
𝑆 = 𝑃 + 𝑗𝑄
Figure 42: Power types table
Real power was the measurement of a circuit’s dissipative elements such as
incandescent lights, ovens, hot water heaters, etc. The units of measurement for
real power were watts and were represented by the symbol (P). Since the power
company only charges residential customers for the usage of real power this
would be the main focus of our power measurements. As noted above in table
42 the equation that governs real power is: 𝑃 = 𝑉 ∗ 𝐼 ∗ cos 𝜃. We know that (V)
was for volts and unless otherwise noted should always be considered to be the
RMS voltage of the given input. The symbol (I) was for current and also should
be considered to be the RMS value unless otherwise noted. The symbol (𝜃) was
known as the impedance angle and was the representation of the amount of
reactive and real power present in a circuit. Figure 43 shown below was a
representation of the power triangle and the relationships between the different
units of power. Once the impedance angle was known the amount of real power
in a circuit could easily be derived with basic trigonometric functions and this was
where the cosine function comes into the equation. [4]
Figure 43: Power triangle
66
The cosine of the impedance angle was directly proportional to the amount of
real power in a circuit, which was called power factor PF. The power factor was
the cosine of the phase angle between voltage and current. A power factor of 1
represents a purely resistive load and power factor less than 1 represents a
reactive circuit. Power factor was also defined as the ratio of true or real power to
apparent power. Power factor takes into account both the phase and wave-shape
contributions to the difference between true and apparent power. In our case,
both the voltage and the current waveforms were sinusoidal. Therefore, to
measure the power factor all we have to do was read the time difference
between zero crossings of the waveform.
The power company wants a unity power factor because if the power factor was
less then unity then the power company had to supply more current for the same
load. Low power factor was caused by inductive loads such as transformers,
electric motors, and high-intensity discharge lighting, which were a major portion
of the power consumed in household. Unlike resistive loads that create heat by
consuming kilowatts, inductive loads require the current to create a magnetic
field, and the magnetic field produces the desired work. The total or apparent
power required by an inductive device was a composite of the real power and the
reactive power discussed above. In order for our power measurement to be
within tolerance, we have to make sure that the error caused by the power factor
was acceptable.
3.2.1 Current Measuring Circuit
Our group decision for the current measuring component for our design was Hall
Effect Current Sensor. To reach this decision we evaluated many options.
Among these options were the Current Shunt, Magneto Resistive Field Sensor,
Hall Effect Current Sensor, Eddy Current Sensor and the Current Transformer.
The Current Shunt was not a good option for several reasons. One was the heat
problem since current going through a resistor in conducting wire produces
voltage and this means at the same time there was a power and there was heat
dissipation and this could be interpreted as the more the current the more the
heat. Heat was something we don’t want in our circuit since it could harm
everything around it starting from the shunt itself. Another reason why we cannot
use current shunt in our circuit was that heat would change the resistance so the
measurements wouldn’t be as accurate with the rise of the current since the
resistance would be a different value and this wouldn’t serve project’s main goal
was to measure power with accuracy. This accuracy may not be achieved if Eddy
Current Sensor was our choice current measuring component neither. In our
power measuring strip design, out of many components more or less all of them
would have magnetic fields since they were all electrical instruments. These
components would interrupt the precision of the Eddy Current Sensor. On the
other hand; Hall Effect Current sensor looks way more superior in measuring the
circuit task because of its advantages such as: High speed and repeatability
67
highly linearity, a very long life and isolation from another high voltage in the
same system, so it wouldn’t be affected by the other components just like Eddy
Current sensor would. Current transformers were a no go as well since they were
really good for higher voltage measurements but their sizes were usually larger
than the our design would allow and they wouldn’t be as beneficial as neither
magneto resistive field sensor nor hall effect current sensor. The superiority of
two sensors above compared to current transformer would be summarized as:
for Hall effect current sensor; logic capability, highly linearity, non mechanical
structure, resistance against heat, perfect to measure high currents for safety
concerns, were insensitive to dust, vibration, humidity, cold and hot and well
sealed structure. On the other hand for magneto resistive field sensor
advantages over current transformer would be: good accuracy and high
sensitivity, durability, fast response rate, decent size, could be used in harsh
conditions and reasonable price.
The two alternatives left, Magneto Resistive Field Sensor and Hall Effect current
sensor were both good choices for our power measuring circuit design but our
choice was towards Hall Effect Current Sensor. Besides the advantages they
would bring to our power measuring circuit design reviewed in the above
paragraph we had to go into detail to compare the two in order to make the best
decision for our design. When two compared, we sow that Magneto resistive
sensor had limited linear range, temperature drift and was sensitive to interfering
magnetic fields from other components. Under these circumstances, our choice
of going with Hall Effect current sensor was justified.
Below in figure 44 was HONEYWELL S&C CSLA2CD; our choice of Hall Effect
Current Sensor as a part of the Power Measuring Circuit Blog Diagram:
SOLID
STATE
RELAY
POWER MEASURING
CIRCUIT
CIRCUIT
CURRENT
MEASUREMENT
MSP430
CSLA2CD
MICRO
CONTROLLER
OUTLET
SOURCE
VOLTAGE
MEASUREMENT
Figure 44: Power measuring circuit block diagram
68
Earlier we were discussing that; in the design part for the solid state power relay,
we decided that SSR should be the first component to be place closest to the
outlet in our circuit design in order to be able to shut the power off whenever we
want from the closest point to the source. The next component in the design was
the Hall Effect current measuring sensor. Since the power measuring strip would
have bunch of electrical devices plugged in to it each drawing different currents,
our Hall Effect current sensor would be measuring the sum of these currents.
Among many products our pick was HONEYWELL S&C CSLA2CD.
As mentioned earlier and could be seen in the blog diagram above Honeywell
CSLA2CD Hall Effect current sensor was the first component of the power
measuring circuit other than the Voltage measuring circuit. The specifications of
the component were as follows:

















Current Measuring Range AC & Dc Up to 72A
Sensor Type Open Loop Linear
Sensed Current Type ac or dc
Sensed Current Range 0 A to 72 A
Package Style PCB Bottom Mount
Output Type Voltage
Sensitivity 32.7 mV N* ± 3.0 @ 8 Vdc
Supply Current 20 mA max.
Offset Voltage Vcc/2 ± 2 %
Supply Voltage 5.4 Vdc to 13.2 Vdc
Offset Shift (%/ °C) ± 0.02
Response Time 3 µs
Operating Temperature Range -25 °C to 85 °C [-13 °F to 185 °F]
Storage Temperature Range -40 °C to 100°C [-40 °F to 212 °F]
Housing Material PBT Polyester
Mounting PCB on 3 pins
Pinout Style 3 pin
In order to power Hall Effect current sensor, we would need an additional circuit.
This circuit had to be a DC circuit and be able to provide 5.4 to 13.2 VDC supply
voltage to power the component. Estimated voltage consumption of the
component should be around 8 VDC at 20 mA.
Right after measuring the current; Hall Effect current sensor would produce Hall
Voltage proportional to the amount of current we were measuring. This voltage
amount would be in AC form and be directed to the analog pin of MSP 430 in
order to be a part of our power consumption calculation. But, of course
compatibility with MSP 430 was imperative since the analog pin input had to be
between 0 to 5 Volts. This means that our Hall voltage had to between 0 to 5
69
Volts so that MSP 430 could start the power calculation. A unique property of
Hall Effect current sensors was the output voltage was very much related with
the input supply voltage to power the component. This means that, whatever
input voltage we chose between the allowed specifications, which were 5.4 to
13.2 VDC the output voltage should be proportional to it. This means that,
according to how we design the supply voltage circuitry of the current measuring
circuit we have to come up with a way to make it be compatible with MSP 430.
An important point to pay attention here was we have to keep in mind that there
would be other electrical components that would need DC supply voltage as well
such as Solid state relay and op-amps, so it would be wise to have a 12 VDC
and distribute it accordingly to necessary components by parallel circuitry with
suitable voltage dividers.
Let’s look at Honeywell CSLA2CD linear current sensors output voltage
characteristics shown in figure 45 according to its supply voltage and then
discuss possible design scenarios. The measured current to voltage output
relation of Honeywell CSLA2CD linear current sensor was shown below
(Assumed supply voltage was 8VDC):
Figure 45: Hall Effect sensor output
Permission requested from http://scienceshareware.com
As seen in the figure above when current was “0” sensor outputs 4V. This was
exactly half of the assumed 8 VDC supply voltage. Another important thing to
know about the component was every 1 A increase in current would cause 0.033
Volts increase in Hall Voltage. This means that in order to calculate the current
the function we have to program into MSP 430 is:
70
MEASURED CURRENT = [VOUT - 4.0] / 0.033
Since we started assuming the supply voltage as 8VDC lets base the first design
scenario to this set up. From the outlet we should expect max delivery of 15
Amps but to be safe lets do our calculations based on 20 A. If we were to draw
20 amps, then according to the formula above the V out which would enter the
MSP 430 analog pin should be [( 20 * 0.033) + 4.0] = 4.66 V which was within 5
V rate that MSP 430 has. In this scenario if the supply voltage was 9 VDC then
final Vout would be 5.16 V and this would be above so it wouldn’t be compatible
with MSP 430. The other optional scenario would be a higher supply voltage let
say 12 VDC. This time Vout turns out to be 6.66 V for the highest possible current
measurement for 20 A even if it was 12 A it would be 6.4 V still too high for MSP
430 to read. In this case we would have to build a op-amp circuit with a gain less
than 1 so the voltage would be regulated. An inverting op-amp would be the right
decision so the output would be a negative voltage and we could add some offset
to it to bring it above 0 volts level. To give an example if the gain of the op-amp
circuit was -0.5, an input Hall voltage of 6.66 V with maximum current possible
would be -3.33 V and min current of 0 would give 6 V which would end up as -3.0
V in the output of the op-amp. If we were to give 5 V offset to this result it would
be between 2 and 1.67. Since the op-amp choice was an inverting one the higher
the current the lower the Vout. In this case 0 A current would give 2 V and 20 A
would give 1.67 V both within the MSP 430 limits.
Out of the two possible scenarios of design the first one seems to be easier to
do, so we decided to go with the first option.
3.2.2 Voltage Measuring Circuit
This part was way easier to design compared to the other parts. Although it was
called the voltage measuring circuit, we already know that it was 120 VAC
(RMS). The point of this circuit design was to represent the voltage sine wave to
MSP 430 through its analog pin, so that we could have a power calculation. As
we could recall from the previous design this analog pin was rated 5 V so this
means some work had to be done to get 120 VAC to 5V. As a matter of the fact
we just have to know about the sine wave of the voltage since we were
interested in the “0” crossings of the sine wave. The reason for this as explained
before was our power formula; P = V * I * cosƟ, and if we recall in order to find
“Ɵ”, we need to know the time difference between the current and voltage waves,
so we have to look at the “0” crossings of both waves. While designing the
voltage sensing circuit, after the voltage divider it would be wise to put an optical
isolator for two good reasons. The first reason is; since this circuit was not limited
to the Hall voltage as in the current measuring circuit, but connected directly to
the grid there might be some voltage spikes and these would harm the micro
controller and opto-isolator would prevent that. Second of all it was easier to be
able to tell the zero crossing point of the voltage wave with this component. Once
we know the voltage signal’s zero crossing time and from the previous current
71
sensing circuit we know the current sine-wave’s zero crossing time, MSP 430
could calculate “Ɵ” and could apply it to; P = V * I * cosƟ formula to find the real
power.
3.3 Power Relay Design
Since our system design was all about managing power and for this reason for
us to manage the power we have to have control over it and this means for us to
be able to turn an outlet on and off which consumes too much power. In order to
this we need a switch which we could control remotely and the electrical switch
component in circuits was called the power relay. There were two main types
mechanical and Solid State Relays as mentioned before in the research part. Our
research resulted with the group decision of going with the Solid State Relays.
Solid State Relays were semiconductor based switches and unlike electro
mechanical relays, they don’t have moving parts which means they don’t need
maintenance unlike changing of the coil was had to be done in contactor relays.
This makes them very reliable with their life time. Solid State Relays were known
for their quick responses and dependable long life and there was no internal
arching that would harm any semi-conductors. Another advantage we would
consider was its PCB compatibility both because of its small size and better
technology. Again solid state relay have the benefit of zero crossing switching
which reduces noise in the as a result the circuit could experience switching
where the voltage crosses were zero. Our group decision was going with the
Photo-Coupled Solid State Relay and our pick, Sharp S0116S01 Solid State
Relay was shown in figure 46 below.
Figure 46: SHARP S116S01 SERIES
72
Here were some specs about our choice Sharp S0116S01 Solid State Relay
shown in figure 47:












Input Forward Current (IF) = 50Ma
Input Reverse Voltage (VR) = 6 V
Output RMS ON-state current (IT(rms)) = 16 A
Output Peak one cycle surge current (Isurge) = 160 A
Output Repetitive peak OFF-state voltage (VDRM) = 400 V
Output Non-Repetitiv peak OFF-state voltage (VDSM) = 600 V
Output Critical rate of rise of ON-state current (dI/dt) = 50 A/µs
Output Operating frequency (f) = 45 ~ 65 Hz
Isolation voltage (VISO(RMS)) = 4.0 kV
Operating temperature (TOPR) = -25 `~ +100 °C
Storage temperature (TSTG) = -30 ~ +100 °C
Soldering temperature (TSOL) = 260 °C
Figure 47: Internal connection diagram
Permission requested from WWW.media.digikey.com
As later would be shown in the main schematics of the whole power measuring
circuit design, we were going to place the solid state relay at the entrance of the
circuit where the current was received from the outlet. The reason for this is;
since this was the entry point and in case of over usage of power we have to cut
the power off from the closest point to the source which was the outlet. This
needs a control unit to monitor the power usage and display it to the user and let
the user to decide to shut the power off. In order to do this, micro-controller
wirelessly connected to a LCD display was planned to be used. Once the read
form the controller was received by and if the user makes the decision to turn the
power strip off, the user interaction from the LCD screen wirelessly transferred to
the microcontroller and from the microcontroller signal goes to the solid state
relay which was connected to one of microcontrollers I/O pins. The compatibility
of Sharp S0116S01 Solid State Relay with MSP 430 through I/O pin works out
73
perfectly since both MSP 430 I/O pin output and Sharp S0116S01 Solid State
Relay’s supply current were the same: 50 mA DC current.
The most important characteristic of the component after the compatibility with
MSP 430, was the output current ceiling As we have discussed in the power
analysis of the power measuring circuit the outlet had a 20A current rating and an
outlet at a North American house hold should be expected to have 12-15 Amps.
This was usually 12 for houses and 15 A for the commercial buildings ready to be
drawn. Since our choice; Sharp S0116S01 Solid State Relay had a rating of 16
Amps we ensure that there wouldn’t be any problems with the drawn current
because 12 to 15 Amps were below 16 which was within the limit.
Another important point we paid attention was the normally open state and the
number of contacts of the solid state relay. Sharp S0116S01 Solid State Relay
was normally open and had four contacts which in other words could be
described as poles. Normally open state would let the current be drawn from the
main grid under normal conditions unless an input signal was received from the
microcontroller to shut the current off in case off an power overage which was
done manually from the LCD screen by the user.
Block diagram for centering solid state power relay was shown below in figure
48, thick arrows represent the current being measured while thin black ones were
the data signals:
SOLID STATE
RELAY
Sharp S0116S01
CURRENT
MESUREMENT
MSP-430
MICRO
CONTROLLER
OUTLET
SOURCE
VOLTAGE
MEASUREMENT
Figure 48: Blog Diagram For SSR
74
The four poles were designed for perfect adaptation for a printed circuit board, it
was very easy to solder and the soldering instructions were shown in figure --. As
mentioned before there were four poles two control input ports and two output
ports. One of the input ports would receive the connection coming from the MSP
430 and the other input port should be grounded. The other left 2 ports which
were the output ports would be used to connect the component in series to the
coming circuit just at the entrance from the outlet, in other words current that
would go to the power measuring components should enter the component from
one output and leave from the other one. This way if the user wants to
disconnect the power going into the power strip this would be achieved.
3.4 MSP430
After careful consideration of all factors with each of the microprocessors above,
our group decided to use the MSP430AFE series from Texas Instruments as our
microprocessor in the power strip module. This particular series of
microcontrollers was specifically designed for the purposes of utility metering and
uses a reduced-instruction set computing (RISC) architecture, meaning that there
was less time between an input and the implementation of that command, taking
into account computations calculated in between. Other benefits for using the
MSP430 were its extreme low power requirements, especially when in sleep
modes, and the wide range of peripherals from TI for electricity metering.
The MSP430AFE253 was the exact microprocessor that would be used in the
power strip portion of our project. “MSP” stands for “Mixed-Signal Processor”;
430 was the MCU platform; F stands for Flash-based memory; E means it was
optimized for e-meter applications and end-equipment; 2 was the 2nd generation,
and 53 was the series in the 2nd generation. The MSP430AFE253 was 7.9mm
long, the body was 4.5mm wide (with the length of the pins added, it was 6.6mm
wide). Each pin was 0.3mm wide, and with 0.65mm spacing in between each pin.
The height of the MSP430 was a maximum of 1.2mm. The dimensions were
shown below pictorially in Figure 49. The top left drawing was a top view; the topright drawing was a magnified view of a single pin; the bottom-left drawing was a
side view of the entire MCU; and the bottom-right drawing was a side view
rotated 90o.
75
Figure 49: MSP430AFE253 dimensions
Permission requested from www.ti.com
Available with this microcontroller was a software suite with a built-in library of
functions to more easily and quickly manipulate data for a variety of purposes.
This library contains a large amount of code support to perform many of the basic
calculations related to and required for energy metering. These values include
those found in figure 50, shown below.
Base
Specific
Voltage
RMS
Peak
Current
RMS
Peak
Energy
Reactive
Active
Apparent
Power Factor
Temperature
Frequency
Tamper detection
Figure 50: Quantities, MSP430
Information from www.ti.com
76
Another library available from TI was the Capacitive Touch Library, which would
help in enabling the capacitive touch capabilities for our MSP430. In the case
that our group chooses to use a touch screen LCD hooked up directly to an
MSP430, this library of code would be very helpful in helping us to learn the
environment as well as write the necessary code for our project. Several
references and guide manuals were also on TI’s website for free download which
help guide first-time users through coding a touch screen interface with the
MSP430. The code for the MSP430 could be written and developed within the
free environment of a program called Code Composer Studio. This program was
free for students, and provides all the support and features needed to code, test,
debug, and run our project using the MSP430.
Another important aspect of the microcontroller was its limitations and
recommended operating conditions. Our group would need to be aware of these
values especially from the PCB implementation stage and through the final
testing, but also when initially designing the circuit diagrams and board layout.
First, the range of temperatures allowed for the device was hardly restrictive, and
should not be a cause for concern for our group; once programmed, the device
cannot be at a temperature lower than -40oC or higher than 85oC. Of greater
concern were the maximum and minimum voltage and current values for each
part of the MCU, as well as the processor frequency range. Values below were
recommended values, not absolute extremes.
Supply voltage, Vcc:
Supply voltage, Vss:
fsystem:
1.8V – 3.6V
2.2V – 3.6V
Vcc=1.8V
4.15 MHz
Vcc=2.7V
9 MHz
Vcc >=3.3V 12 MHz
Additionally, the recommended range of system frequency (f system) changes
throughout the programming process, as well as during normal running
operations. If possible, the supply voltage should not exceed 3.3 V during
execution of the program, as shown below in Figure 51.
77
A)
B)
Minimum processor frequency was defined by system clock. Flash
program or erase operations require a minimum Vcc of 2.2 V.
If high frequency crystal used was above 12 Mhz and selected to
source CPU clock then MCLK divider should be programmed
appropriately to run CPU below 8 MHz.
Figure 51: Frequency range for MSP430
Permission requested from www.ti.com
The clock system of the MSP430 was designed to allow for some degree of
flexibility. By choosing one of the three clocks and setting the clock speed as low
as possible, the power consumption could be reduced. The first clock was the
LFXTCLK (Low Frequency Crystal Clock), which connects to the XIN and XOUT
pins and had an oscillation of 32kHz. The second clock, the XT2CLK (Crystal 2
Clock), connected to the XT2IN and XT2OUT pins, was intended to be the highspeed clock source. Finally, the third clock, the DCOCLK (Digitally Controlled
Oscillator Clock), was an internally generated clock source that runs at 900kHz
(which could be reduced internal using clock dividers). Essentially, when an
ultralow-power stand-by mode was desired, the low-frequency auxiliary clock
(LFXTCLK) should be used; for high performance signal processing, the highspeed master clock (XT2CLK) should be used.
The interrupt vectors, which each contain the corresponding 16-bit address of the
code for the interrupt instruction, were located from address 0FFE016 to 0FFFF16.
In this range was also the power-up starting addresses. These addresses should
be known to our group and were shown below in Figure 52.
78
INTERRUPT
SOURCE
Power-up
External reset
Watchdog
Flash memory
INTERRUPT
FLAG
WDTIFG
KEYV
SYSTEM
INTERRUPT
Reset
NMI
Oscillator fault
Flash memory
access
violation
Watchdog
timer
Timer_A3
NMIFG
OFIFG
ACCVIFG
Non-maskable 0FFFE16
14
WDTIFG
Maskable
0FFF416
10
TACCR0
Maskable
CCIFG
TACCR1 and Maskable
TACCR2
CCIFGs,
TAIFG
URXIFG0
Maskable
0FFF216
9
0FFF016
8
0FFEE16
7
UTXIFG0
Maskable
0FFEC16
6
ADC10IFG
Maskable
P2IFG.0
to Maskable
P2IFG.7
P1IFG.0
to Maskable
P1IFG.7
0FFEA16
0FFE616
5
3
0FFE416
2
Timer_A3
USART0
receive
USART0
transit
ADC10
I/O Port P2
I/O Port P1
WORD
ADDRESS
0FFFE16
PRIORITY
15 (highest)
Figure 52: Interrupt address locations
Reproduced with permission requested from www.ti.com
The flash memory in the MSP430 could be programmed in one of two ways:
using the in-system process by the CPU, or via the Spy-Bi-Wire/JTAG port. Most
locations could be erased one at a time, or the entire 0 to n locations (512 bytes
each) could be erased in one step. Words were organized with the low byte
placed at an even address and the high byte placed at the next highest odd
address (e.g. LB=000416 and HB=000516). As shown below in Figure 53, there
were separate address spaces for different types of memory and registers.
Important to us were the Flash/ROM memory spaces (1000016-1FFFF16), the
Interrupt Vector Table (0FFE016-0FFFF16), and the Special Function Registers
(016-0F16). Also, any 16-bit peripheral modules were mapped into the address
space 01016 to 0FF16. Code could only be accessed on even addresses, and
data may be accessed as words or bytes. [8]
79
Figure 53: Address space of MSP430
Permission requested from www.ti.com
3.4.1 Programming the MSP430
Before writing code for the microcontroller, it was helpful to try and understand
the physical architecture of the chip as well as the interconnections between
each of the pins and internal devices. The entire functional block diagram was
shown below in Figure 54. First, the clock system had an input port (XT2IN) and
an output port (XT2OUT), both of which double as digital I/O pins. From here, the
clock was of course communicating to the CPU. Next, the 8-wide digital I/O port
P1.x communicates with the Timer_A3 as well as the main CPU, via MDB.
80
Figure 54: MSP430AFE253 Function Block Diagram
Permission requested from www.ti.com
The MSP430 could be programmed in one of two ways: using the in-system CPU
process, or externally using the JTAG/Spy-Bi-Wire Port, which uses flash
memory. Programming through the JTAG interface requires four input signals
and one output signal, in addition to ground and Vcc, if powered externally. The
input pins were TMS, TCK, TDI, and TEST, and the output was TDO; the pins
were summarized below in Figure 55, along with their purposes.
Figure 55: JTAG Interface Signals
Permission requested from www.ti.com
81
The terminals and I/O pins and their locations on the microcontroller were shown
below in Figure 56. As seen below, the two analog input pins (A0 and A1) were
located in the upper left corner of the microcontroller, with the analog supply
voltage located directly below them, and a third analog input pin located beneath
that. The higher pin in each pair was the positive terminal (pins 1, 3, 5, and 8),
and the lower pin was the negative terminal (pins 2, 4, 6, and 9). The last three
pins on the left hand side of the microcontroller were TEST, (RST)’, and the first
general purpose I/O pin (pins 10, 11, and 12). The right-hand side of the
microcontroller consists primarily of general-purpose digital I/O pins (pins 14, 15,
and 17-24), in addition to a digital supply voltage (pins 13 and 16).
Figure 56: MSP430AFE23 Pin Designation
Permission requested from www.ti.com
In addition to a detailed diagram of the exact locations of each pin on the
MPS430, a useful tool would be a chart detailing the purpose, functionality, and
I/O status of each pin. This diagram was shown below, in Figure 57.
82
Terminal
name
A0.0+
A0.0A1.0+
A1.0AVCC
Pin number
I/O capability
1
2
3
4
5
I
I
I
I
N/A
AVSS
6
N/A
VREF
7
I/O
A2.0+
A2.0TEST/
SBWTCK
8
9
10
I
I
I
(RST)’/NMI/
SBWTDIO
P1.0/SVSIN/
TACLK/
SMCLK/TA2
DVSS
11
I
12
I/O
13
N/A
P2.6/XT2IN
14
I/O
P2.7/XT2OUT
15
I/O
DVCC
16
N/A
P1.1-1.7, 2.0
17-24
I/O
Description
Positive analog input (A0.0)
Negative analog input (A0.0)
Positive analog input (A1.0)
Negative analog input (A1.0)
Analog supply voltage,
positive terminal. NOTE:
Must not power up prior to
DVCC.
Analog supply voltage,
negative terminal.
Input for an external
reference voltage/ output for
internal reference voltage.
Positive analog input (A2.0)
Negative analog input (A2.0)
Selects test mode for JTAG
pins on P1.5 to P1.7 and
P2.0.
Reset or nonmaskable
interrupt input.
General-purpose digital I/O
pin.
Digital supply voltage,
negative terminal.
Input terminal of crystal
oscillator. General-purpose
digital I/O pin.
Output terminal of crystal
oscillator. General-purpose
digital I/O pin.
Digital supply voltage,
positive terminal.
General-purpose digital I/O
pin.
Figure 57: Terminal functions of MSP430
Reproduced from www.ti.com, permission requested.
One of the benefits of using this microprocessor was the existence of two allpurpose digital I/O ports. Port P1 includes pins 12 and 17-23, and was 8 bits; port
P2 was 3 bits, and encompasses pins 24, 14, and 15 (both in that order).
Properties of these include:
83




Each of the 11 I/O pins was independently programmable, allowing for any
combination of input, output, and interrupt conditions.
Edge-selectable interrupt was possible on all 11 bits.
Each pin had an individually programmable pull-up/pull-down resistor.
Finally, the unused bits in port P2 (i.e. bits [5:1]) were read as 0, and any
write instructions to these bits were ignored.

Several other features of the MSP430AFE253 were worth mentioning at this
time. First, the MSP430 had a watchdog timer (WDT+) which was intended to
perform a controlled system restart after a software problem occurs. The time
interval chosen for this function could be chosen by the user. In addition, if the
WDT+ was not needed in a program, it may be used as an “interval timer”, which
could provide a system interrupt at predetermined periods in time. This function
may be used by our group to periodically (though at regular intervals) provide an
interrupt to the current and voltage measuring system in the power strip to force
the system to send the data to the microcontroller in the LCD.
A second feature was the brownout circuit/supply voltage supervisor. The
brownout circuit was used to provide a reset signal during power on and power
off periods. The SVS (supply voltage supervisor) circuit was used as a voltage
detection system. If the supply voltage to the microcontroller drops below a
predetermined level (set by the user), this was detected by the SVS. In addition,
the circuitry here was capable of performing both automatic resets and nonautomatic resets of the system. Only after the brownout circuit releases the
device would the CPU begin code execution.
Available on the website for the MSP430 was a PDF file which details the
implementation of a single-phase electronic watt-hour meter using the
MSP430AFE2xx series of microcontrollers. This documentation was a valuable
resource for our group and would be used as a cross-checking for our group at
various stages throughout the months to ensure that our plan for implementation
was viable. Part of this report includes details on the hardware implementation
used in the watt-hour meter, including the two power supplies used. The first was
a resistor capacitor power supply, shown below in Figure 58. The circuit uses a
series of devices at the input in an effort to reduce the dc voltage provided to the
MSP430 to the appropriate level of 3.3V for full-speed operation. Additional help
and support, as well as the design equations for the power supply, were shown in
the ‘Capacitor Power Supplies’ section of MSP430 Family Mixed-Signal
Microcontroller Application Reports. The left-hand side of the diagram comes
from a 110/220 VRMS alternating current supply source, and the right-hand side
goes directly into the digital supply port and the digital ground port of the
MSP430AFE.
84
Figure 58: MSP430 Energy Meter Capacitive Power Supply
Permission requested from www.ti.com
A second power supply circuit was constructed to ensure that enough power and
current would be available to power RF transceivers, which our project would
also use. Although the microcontroller used in the paper from TI was different
than that which our group would be using, we would anticipate similar issues with
providing sufficient power to the RF transceiver. This power supply would provide
3.3V (DC) from the 110 or 220 V (AC) in order to power the transceiver properly.
[8]
3.4.2 Code Composer Studio
This section would detail the process for programming the MSP430, and would
include all pertinent information related to programming. The programming of the
MSP430 could be done entirely in the C language, and was done in the Code
Composer Studio (CCSv4) environment. CCS was a whole suite of tools
including compilers for each of the MCUs from TI, a debugger, source code
editor, project build environment, and simulators, and had been made available
for free to the engineering students at the University of Central Florida. To begin
the coding process, our group needed to become familiar with the programming
process for the MSP430G2231 microprocessor included in with the Launchpad
development kit. These step-by-step instructions would be used every time that
our group begins to code a new section of our project; the only difference would
be the microprocessor that would be chosen (see step 2 below), which would be
the MSP430AFE253. The process was detailed below:
1)
2)
3)
4)
Open Code Composer Studio and select OK to choose the default
workspace.
Choose “MSP430” as the Project Type, then after two pages, select
MSP430G2231 (MSP430AFE253 later on) as the Device Variant.
At this point, you could open a new project by selecting File->New->New
Project.
The window main.c was where the code could be written. You could also
link an existing c file to your program by navigating Project->Link Files to
Active Project, and then navigating to the file(s) needed. This was useful
85
in linking files which contain the regimen for calculating voltage values,
multiplying two numbers, or a number of other small operations which
could be referenced.
The coding process involves using many of the functions provided with the
software to manipulate the data flowing into the MSP430. The following
functions, shown in Figure 59 below, would likely be used in one way or another;
this list would be used as a quick reference over the next several months while
our group edits and debugs our code.
Function
name
dc_filter
Action
debounce
Used to debounce a push
N/A
button switch, i.e. eliminates
errors associated with
pushing a button to activate a
clock edge.
Shifts a 48 bit integer
32
upwards by a specified
amount, and then divides a
16 bit integer into the 48 bit
integer.
div_sh48
div48
imul16
isqrt32
mul48
q1_15_mul
Filters the DC content from
an AC mains waveform
signal.
Divides a 16 bit integer into a
48 bit integer.
Completes a 16x16 2’s
complement multiplication.
Computes the square root of
a 32 bit number, where the
last 16 bits were after the
decimal point. NOTE: be
careful not to have a
negative number as input.
Computes a 32x16
multiplication function.
Computes a 16x16
multiplication function.
86
Bits of
answer?
Same as
input
32
32 2’s
compleme
nt
16
32
32
Use in project
Filters out DC
content of Voltage
and Current for
Power calculations,
leaving only AC.
Used for testing and
debugging the
circuit (i.e. a manual
clock edge).
Combines two
operations into one
to increase
efficiency and
decrease time
spent. Useful for dot
product calculations.
For dot product
calculations.
Multiplication of
Current and
Voltage.
Figure 59: Table of functions for MSP430 programming
The process for our data collection code was detailed below in the flow charts in
Figures 001 and 002 below. Our code begins with the connection of the power
sources to our microprocessor, then an initial system reset. The next step was
the hardware setup, including: analog-to-digital converter, digital general-use I/O
pins, Clock system, timer, and Universal serial asynchronous receiver/transmitter
(USART). This initial process was shown in Figure 60.
Figure 60: Initial system setup flowchart
Next, the system checks if the main power was on. If not, the device would go
into LPM0, and then wakeup. With the main power on, the system would check
for a signal from the power strip module to ensure proper connectivity. Once this
was passed, the energy accumulation phase could begin as seen in figure 61.
First, a small amount of time (perhaps 1s) of data was accumulated. Then, a
check in the system was made to see if enough data had been gathered for
calculations of current, voltage, power, etc. If not, the system returns to the
previous step and gathers another second’s worth of data. When a sufficient
amount of information was stored, the system would move to the next step and
calculate the RMS values of the parameters needed. These would be stored and
87
then sent to the wireless transceiver through the SPI/UART (Serial Peripheral
Interface/Universal Asynchronous Receiver/Transmitter). After transmission of
the data, the system would return to the beginning of the program (but after the
initial setup of the clock, port pins, etc) and perform a system check for the main
power. The process would repeat indefinitely until the power was disconnected or
an interrupt signal was received from the hub microprocessor in the LCD.
Figure 61: Looping data collection/transmission process
3.5 Wireless Design
The wireless design that follows shows a visual representation of the data flow
between devices. The CC2500 retrieves input from the main MSP430 controller
and transfers it to transceiver on the power strip. This power strip only had to
send data out; it does not take in any data. The only inputs were the awake and
control signals. The awake signal takes the MSP430 out of sleep mode, when it
receives this signal was also transfers all of its data measurements to the
88
CC2500 and then returns to sleep mode. The control lines shown in figure 62
send a word of information containing the information to turn devices on or off.
The control signal also takes the MSP430 out of its hibernation state. It was
important to note that the transceiver does not send any data regarding the sleep
mode. This was done locally through the MCU software.
Data Out
CC2500
CC2500
Awake
Contr
ol
Contr
ol
Data Out
Contr
ol
Data Out
Awake
Awake
LOCAL
MSP430
Power Strip
MAIN
MSP430
Control
Figure 61: Wireless Data Flow
3.5.1 ATS Wireless Module
The automatic transfer switch must have the capability to wirelessly transmit
whether it was in normal or emergency mode back to the touch screen controller.
The automatic transfer switch does not come standard with this module so we
would have to design it ourselves. The design for this module was rather simple
and was not considered to be part of the “main design” of the project. The
reason for the adding the automatic transfer switch to the project was for the
purpose of demonstration. Without being able to simulate the transfer between
normal and emergency power the project would not be able to show its
89
usefulness. The project would simply look like any other home power monitoring
system, which was not the intent of our groups design.
The automatic transfer switch already provides a means of signaling whether it
was utilizing normal or emergency power via low voltage signal so all we have to
do was send that signal back to the touch screen controller. We were going to
use the same TI CC2500 wireless chip that would be utilized in the power strip
modules and touch screen so adding it to the network would be without much
effort. The rating of the CC2550 interrupt input rating was 3.3 volts and the
transfer switch’s signaling circuit outputs 12 volts so a voltage dividing circuit
would need to be added to make the two compatible. In addition the CC2500
requires a DC power supply between 1.8 and 3.6 volts and we would utilize a
battery to supply this voltage. Because of the ease of removing the service
compartment cover of the transfer switch and access to terminals 5 and 6 we
believe that making a small module with battery power would be the smallest and
most economical option to us. It would be extremely easy for a homeowner to
change the battery when it was time to be replaced. Below in figure 62 a simple
block diagram of the automatic transfer switch module was shown.
Figure 62: ATS Block Diagram
90
3.6 Generator Module
During times of power loss the home generator becomes the lifeline of the homes
power system. In order to ensure the safeguarding of the generators operating
conditions we must design a module to measure the generators instantaneous
power consumption and have that information wirelessly transmitted back to the
touch screen controller for program manipulations. The output from a typical
home generator was 240 volts with a neutral. The ground was picked up from
the houses existing grounding system.
The power measurement module for the generator would be slightly different
than that of the power strip modules because we have two phases to monitor
instead of one. This was not a problem though as we only need one additional
power measuring circuit to accommodate for the extra phase. The total power of
the generator was simply the addition of the two individual phases. The following
equations govern the total power consumption of the generator.
𝑃𝑃𝐻𝐴𝑆𝐸 𝐴 = 𝑉𝐴 ∗ 𝐼𝐴 ∗ cos 𝜃
𝑃𝑃𝐻𝐴𝑆𝐸 𝐵 = 𝑉𝐵 ∗ 𝐼𝐵 ∗ cos 𝜃
𝑃𝑇𝑂𝑇𝐴𝐿 = 𝑃𝐴 + 𝑃𝐵
Except with the addition of the extra power measuring circuit the main
components remain the same as the power strip modules. First the power would
be measured via current and voltage sensors and then power calculations would
be done using the MSP 430. This information would then be wirelessly
transmitted back to the touch screen controller for further processing and display
to the user. A block diagram of the generator module could be seen below in
figure 63.
91
GENERATOR
PHASE
A
PHASE
B
POWER
MEASURIN
G CICUIT
POWER
MEASURIN
G CICUIT
BATTERY
POWER
CC2500
WIRELESS
MSP 430
QSREEN
CONTROLLE
R
Figure 63: Generator Module Diagram
3.6.1 Generator Power Measurements
In case of a power shortage our ATS module was going to replace the power grid
with the generator output. There was couple of slight differences between the
regular power measurements and the generator power measurements. The
generator power measuring circuit was going to be limited power output and
there was need to shut it off because of over usage. Because we have to do a
presentation of our project indoors and it would be loud to bring a generator
indoors we would represent generator output with another outlet which we would
limit its output on purpose.
92
Energy in a home type generator was produced in AC form since these types of
generators were just a smaller version of the power plant generators. Usually the
generator was a stand by generator which was connected to the house wiring
with an ATS or it had a plug on it if it was a portable generator. Doesn’t matter
which type it was sometimes it would be 240 VAC but most of the time we’ll
receive in a regular house outlet, 120 with 12 amps of max current could be
drawn. In a regular home there were different appliances with different uses most
of which were resistive loads since they’re used somehow to produce heat like
oven or water heater, besides resistive loads we have Inductive and capacitive
loads as well which have magnetic field outcomes when connected to the
current. . The same way as it was with the power grid, with generator the
outcomes of these two different kinds of loads give different types of power which
were “Real Power” and “Reactive Power”. Our concern was with the Real power
the same way when it was with measuring the power from the main grid because
it was the generic scale of power measurements, so our measurements would be
based on Real Power again.
Since our generator would be producing Apparent power which was measured in
VA ( S = P + jQ ) and as we mentioned before the types of power could be listed
as Real power which was measured in Watts, Reactive power which was
measured in VAR’s. Since the generic power measurement was Real power,
then our power measuring circuit had to be measuring the real power for the
generator as well. This brings us back to the power factor calculation or cosine of
“Theta”. This was the angle between both the voltage and current values or the
same way the same angle between apparent power and real power. According to
the previously mentioned formula “P = V * I * cosƟ” where Theta was the phase
angle all we need to do was multiply the values we get from measuring the circuit
with cosine Theta and we find the power. Calculation of Theta from the voltage
and current measuring circuits would be as follows: All we have to do was read
the time difference between zero crossings of the waveform of both voltage and
current waves. This difference was proportional to the whole cycle period. Since
all these calculations would be done by our microcontroller “MSP 430”. and there
was an allowed range of inputs to MSP 430 by implementing voltage divider
circuit, to both current and voltage measuring we could make these values
compatible with the microcontrollers pin input values.
This limited power output of generator’s would be measured with the exact same
methods that were described in the previous sections, as a matter of the fact it
would be the same power measuring strip that we designed to measure the
power from the regular power grid.
Previously we discussed about how we would design power strips that we would
plug into outlets and these power strips would measure the power consumption
of appliances which were plugged into it and would report wirelessly to LCD
display. Now for the generator output nothing had changed since ATS would
switch to generator power and the power strip would measure the power
consumption of the same appliances plugged into it and would report wirelessly
93
to LCD display. Since these power measuring strips were plugged into outlets
they don’t care which source was the provider of the power, they just measure it,
doesn’t matter if it was the power from the main grid or the generator.
Once the measured power was sent from MSP 430 to LCD through the wireless
module, this was where things change a little bit. Now we do have a read
reported wirelessly from the power strips but this time the source was not the grid
but the generator. The only way for our main processor to be able to tell the
difference of this change in the power sources was if it receives a message from
the ATS module. As mentioned before ATS was the switch which would detect
the shortage and switch to generator power. In this case either wirelessly or a
hard wire from the ATS module to the main processor would solve this problem.
This way our system would adjust to the emergency protocol and turn on only the
imperative outlets, and turn off all the others in order not to overwhelm the
generator. Another solution would be instead of having ATS and the main
processor talk to each other we could solve the problem with software. This
solution would be if the system overall power consumption falls under a certain
limit, which would be usually the generators power rating we could get the
system override the current status and get into the emergency status. Only
problem with this solution would be the transition period would be overwhelming
on the generator although it was for a short time, but it would save us plenty time
from the design part since it only appears to be an if statement that we have to
add into the code.
3.7 Parts List
The parts list below contains a list of all the components required for our project.
Each appliance controller would consist of one CC2500 Transceiver, one
MSPR430, an outlet, and a circuit board. These components would be housed
separately. The LCD display and the automatic transfer switch were standalone
devices and would be purchased individually. These two devices would not be
designed; they were simply to be interfaced with the rest of the parts of the
Power Safe system. The power metering circuit would also contain the same
components as the appliance controller. In addition it would contain a relay that
could support 30A of current delivered by the generator.








4 CC2500 Transceivers
1 LCD Display
1 Automatic Transfer Switch
2 Housings for outlet/controller
3 current transformers
4 MSP430F2274
5 circuit boards
EagleCAD Software
94
Part Name
Quantity
Cost
Total
Microcontroller Components
~
$10
$20.00
LCD Display
1
$400.00
$400.00
Automatic Transfer Switch
1
Free
Free
Outlet Housing
1
$10.00
$10.00
Current Transformer
4
$8.00
$32.00
MSP430F2274
4
$2.75
$10.00
Initial PCB design
1
$33.00
$33.00
Circuit Board
4
$33.00
$132.00
Solid State Relay
4
$10.00
$40.00
EagleCAD Software
1
$49
$49
Power Outlet
4
$10
$40
$766.00
Total
Figure 64: Parts Prices
3.8 Schematics
The following three diagrams detail the overall schematic for our project. The first
diagram, figure 65 shown below, was the connection of the TI CC2500 wireless
transceiver to the QScreen controller utilizing pins 1 and 2 on Port E.
95
Figure 65: QScreen Wireless Transceiver
The next diagram, figure 66 below, was the complete schematic of: the current
transformer, potential transformer, power relay, microcontroller and wireless
transceiver. These items wired together control the duplex receptacle.
96
Figure 66: Power strip module schematic
The last schematic, below in figure 67, was the current and potential
transformers measuring the current power consumption of the generator.
97
Figure 67: Generator Module Schematic
98
Section 4: Testing
4.1 Procedures
The software for the wireless transmission of data would be developed using the
EZ430-RF2500 Development Tool. Through using the debugging software
provided with this tool we would test to make sure the data being received by the
transceiver was correct. The LCD panel would be tested along with one
development board. Once we have proper communication between these two
devices then the rest of the power strips could be added to test functionality
between multiple devices. All of the initial testing would be done using
breadboards and pre-assembled microcontrollers. We cannot proceed with the
final testing design until all components were working as described.
Once this initial testing had been completed and was working to the correct
specifications we would develop the first PCB. Once again we must test the PCB
with our original design to confirm that no abnormalities have occurred. Once the
first PCB was working we would develop the remaining boards to complete the
project.
Following was the testing procedure we would apply for the project. This test
procedure was developed to validate all of the steps covered under the
requirements and specifications. Following successful completion of the testing
procedure all of the individual parts of this project would be functional. The final
testing procedure would test the final case in which all of the components were
working together. Once this step was complete all of the components would be
individually mounted on their own circuit boards. By isolating the initial testing
procedure with the actual implementation of the printed circuit boards we would
be able to verify that any abnormalities were not a design problem. Actual
results for all testing procedures were yet to be determined.
4.1.1 Power Sensing Circuit
Case Name: Case one – Power Sensing Circuit
Unit tested: Current & Voltage Sensing
Circuit
Assumptions: Test lamp had an expected power usage of 60 watts.
Test Data: Current, Voltage, and Power
Steps to be executed:
99
1.
2.
3.
4.
5.
6.
Plug in a regular 120V AC 60Watt desk lamp with an on/off switch into an
outlet
Measure the current being drawn by the lamp with a multimeter
Calculate the total power consumption and compare the expected values
Plug in the same 60 Watt desk lamp into our circuit
Measure the output current and voltage from our circuit
Compare these values to the actual values from earlier. The values must
be accurate to .1A and 1V.
Expected Result: The actual values measured from the unit should be close to
the advertised power consumption of our lamp. The measured values would be
calculated using the power formulas. The expected power consumed by the lamp
was 60 Watts.
Summary: The power sensing circuit was simply a step down in voltage and
current output. These outputs should be in the range of values that could be
read by the MSP430 ADC10 pins. Given a voltage input of 8 volts and a
resolution of 1024 each step would have an accuracy of around .008 volts. This
means that with each step up or down of .008 volts we could output a new
voltage reading from the power strip to the main control of the home.
4.1.2 MSP430
Case Name: Case two – MCU
Unit tested: MSP430 Microcontroller
Assumptions: Test lamp had an expected
power usage of 60 watts.
Test Data: Current, Voltage, and Power
Steps to be executed:
1.
Take the results of the measured values from test case 1. Calculate the
actual power usage through the power triangle equations.
2.
Physically wire the output of the power sensing circuit into the input of the
MSP430 ADC10 pins.
3.
Through the software read and display the individual current and voltage
values of the power sensing circuit.
4.
Calculate power measurement using MSP430 functions.
5.
Verify that the values calculated from case 1 were equal to the function
calculated values on the MSP430.
6.
Enter sleeping mode
7.
Leave sleeping mode when an interrupt was received
100
Expected Result: The results of test case two should be identical to test case 1.
A power consumption reading approximately 60 Watts was expected. After the
microcontroller sends an information packet it should enter sleeping mode until it
receives a turn-on signal.
Summary: Each microcontroller would be tested individually before they were
interfaced wirelessly. To meet our specifications each microcontroller would
measure current from a controlled target device. Readings from each device
would be measured to make sure they were accurate. Once all of the devices
were individually tested then their data would be sent wirelessly to the main
microcontroller. This information would be aggregated and then transmitted to
be displayed on the LCD panel.
If all the devices were able to measure current and display these measurements
to a screen then the next step would be to transmit a signal from the main
microcontroller back to each power sensing microcontroller. The transmitted
signal would need to accomplish three tasks. Each power strip containing an
MCU must restrict power consumption which would be represented by a light.
Another signal should resume the power consumption. If both of the tasks were
complete then the MCU must also go to sleeping mode while there was no power
being consumed by the power strip. The main microcontroller must have
constant communication with all power strips in the home as well as the MCU
measuring current on the secondary power supply.
4.1.3 Wireless
Case Name: Case three – Wireless
Unit tested: CC2500 Transceiver
Assumptions: The MSP430 MCU was able
to read and output power consumption
information to the transceiver from a given outlet.
Test Data: Hard coded data from software and power measurements given by
the MSP430.
1.
2.
3.
4.
5.
Steps to be executed:
Hard code data into the EZ430-RF2500 wireless Development Tool
Using this code send a data packet to a different transceiver
Read data from the transceiver and compare it to the original packet
Autonomously send measured data from one power strip to another. Each
power strip contains one microcontroller and one transceiver.
Test low power consumption using wireless information. When a sleep mode
signal was received the MSP430 should enter sleep mode. If a turn-on signal
101
was received the MSP430 should send a second packet of information. The
MSP430 should send a packet of information every 5 seconds.
6. Repeat steps 1 through 4 indoors with a line of sight range from 10ft to 100ft.
Again test for this range between 2 to 3 rooms on the same floor and then
between rooms from adjacent floors.
Expected Result: The transceiver should be able to send and receive accurate
packet information between rooms.
Summary: The wireless modules would need to be checked first for functionality
and then for range and reliability. Communication between two to three rooms at
50 feet was necessary.
If adequate range was not met then considering
different antenna options could increase the range. Each unit should be able to
consistently receive and transfer data bytes. Once all power strips were
functional then the design would be tested to find the limitations in range of the
design. The open air tests would be used to determine to maximum range of our
design. By adding walls in between the wireless communication we could see
how much of an effect different materials have on the range of the system.
Case Name: Case four – LCD
Unit tested: QSscreen Controller
Assumptions: MSP430 was functional and
could correctly send and receive data to
other microcontrollers.
Test Data: Numbers generated by MSP430
Steps to be executed:
1.
Cycle through different menu options by touching the LCD screen
2.
Output total power consumption reading information from the MSP430
onto the screen
3.
Send an interrupt control to the microcontroller for powering outlets on and
off
4.
Allow a user to input the current cost per kilowatt hour given by their local
power company
5.
Display current date and time
6.
Cycle through each power strip to read individual power measurements
7.
LCD should display an individual power strips current power consumption
as well as total power consumption
8.
Display total monthly cost for all power strips individually and summed
together
9.
Ability to store and compare the cost of up to 12 consecutive months of
information
102
10.
Calculate kilowatt hour conversions using computer software and compare
the results to the displayed information on the LCD
Expected Result: LCD screen touch screen function would properly cycle through
menus. All of the power reading information would be calculated locally by the
main microcontroller. The LCD process should take this information as input and
apply its own calculations to determine the monthly cost for the appliances.
Power to kilowatt conversions should be determined by the LCD processor.
Summary: The first test for the LCD would be a basic touch test. This test
consists of touching the screen to see if it responds correctly. Touch response
should be speedy and accurate. While touching the screen all the different
menus would need to be tested. This would ensure the boundaries of the
buttons were correct and that the LCD screen does not have any hardware
issues. The LCD unit also needs to be tested for numeric accuracy. Data
received by the microcontroller needs to be displayed on the LCD correctly. All
numbers would be displayed in real time. The LCD unit contains a local
processor that should receive inputted cost per kilowatt hour information and
output total cost for the household.
4.1.4 Software
Case Name: Case five – Software
Unit tested: Software
Assumptions: Microcontrollers could read
information from the power sensing circuit.
Test Data: All data transmitted and received by the microcontrollers.
includes interrupts as well as voltage and current readings.
This
Steps to be executed:
1.
Test the software with a single power strip module
2.
Verify that interrupts were working as described
3.
Output all numeric data to a screen and verify the numbers were correct
4.
Plug in a second power strip and test the software for module priority
Expected Result: The code should display all data correctly and the priority
routines should work as described. If a priority power strip exceeds the power
distribution of the backup generator then it should turn off immediately.
Summary: Software would be tested throughout the testing process. Component
specific tests would reveal any abnormalities with the software. After the home
modeling system was complete a full function test of the code could be
completed. The software must be tested to see if it was displaying the correct
103
information. All of the math must be verified within the code so no measured
values were incorrectly compared. Non essential features in the coding of the
LCD would be tested last. The LCD should display all of the necessary
information in the correct location. The code must not present any abnormalities
when a user requests information from the LCD. The code should also be tested
for speed and efficiency.
4.1.5 Automatic Transfer Switch
Case Name: Case six – Automatic Transfer Switch
Unit tested: ATS
Assumptions: The transfer switch was
simply transferring power from one power supply to another.
Test Data: Current
Steps to be executed:
1.
Interface the ATS with one primary 120V outlet
2.
Read the output current from the ATS
3.
Interface a second 120V outlet
4.
Remove the primary power source from the outlet
5.
Measure the current coming out of the ATS
6.
If there was current after step 5 then the ATS was functional
Expected Result: The automatic transfer switch would autonomously switch
power output from the primary to the secondary power supply once the primary
power was lost. Power delivery from the ATS should not be interrupted.
Summary: The primary focus of the project was not on the switch. The automatic
transfer switch was a self contained and individually manufactured part of our
project. The only test necessary was to ensure that the device was functional. If
the device was functional it could be implemented into the project.
4.1.6 Scaled Test
Case Name: Case seven – Scaled Test
Unit tested: All Units Simultaneously
Assumptions: All units have been tested
individually and were working according to the specifications and requirements
section of the project. Before this test was conducted steps 1 through 5 should
receive a passing grade.
104
Test Data: Voltage, Current, Power from different appliances and different outlets
simultaneously, digital interrupts would also be tested.
Steps to be executed:
1.
Wire up all circuits and plug in the power source for the ATS
2.
Turn on the generator and remove the primary source of power from the
outlet.
3.
Observe circuit for any abnormalities
4.
Send interrupt and control inputs to all of the MS430 boards
5.
Read information on the LCD
6.
Cycle through all menus on the LCD and test all menu functions including
the ability to set priority devices.
7.
Plug appliances into each outlet to verify the power strips were functional
and that the priority power strips remain on if power drain exceeds power
delivery.
Expected Result: Circuit should behave normally and all components should
work as they were individually tested before.
Summary: The scaled modeling of the home monitoring system consists of three
power strips, one controlling station with an LCD, and a monitoring device on the
secondary power supply. This was our final design implemented using
breadboards and microcontrollers. After it passes this testing phase then the
circuit would be implemented using PCB technology. The project would be
demonstrated using two sources of power. One source would be directly
connected to a conventional outlet supplying 120V of AC current. Another
source would provide the system with a limited and varying amount of current.
The secondary source would be represented by a small generator. By altering
the current supplied different outlets would be shut off according to their priority.
Outlets which have higher priority would remain turned on if there was sufficient
power to supply them. All of the information regarding each outlet and the
secondary power supply load would be displayed on the LCD. Users would have
the ability to interact with the LCD to see the intended information displayed
about each power strip. All functionalities would be tested with respect to the
specifications and requirements. A visual representation of the final scaled test
was given below. Communication data flow between all of the devices was show
by the arrows.
Section 5: Administrative Content
5.1 Timeline
June 8, 2011
 Decide on the final version of the specifications of the project
105

Finish and submit initial project and group identification document
June 15, 2011
 Research the parts needed for the hardware parts and microcontroller of
the project
 Discuss and research the compatibility of these parts with each other
along with their price, availability and delivery timing
June 22, 2011
 Organize parts list according to parts’ availability and vulnerability and add
back up parts to this list and finalize the list
 Finalize parts list
 Order supplies
June 29, 2011
 Double check and confirm the delivery of the parts
 Start mastering the coding necessary for the micro controller
 Start the Senior design paper
July 6, 2011
 Start code draft implementation
 Discuss each group member’s progress on final paper on weekly
meetings
July 13, 2011
 Discuss each group member’s progress on final paper on weekly
meetings
 Half-way point where each member was expected to have written 15
pages
 Proofread, discuss, and edit each other’s work so far completed
 Prepare for the Team Presentation
July 20, 2011
 Prepare for the Team Presentation
 Discuss each group member’s progress on final paper on weekly
meetings
July 27, 2011
 Team ready for Presentation
 Complete Senior Design paper
 Last editing of the senior design paper
August 3, 2011
 Submit senior Design paper
 Present the project to class
August 10, 2011
 Continued research over semester break
106

PCB Eagle file reviewed and edited
August 17, 2011
 Possibly consider potential minor changes of parts for various reasons
found in research over break
 Small segments of code should be tested on the MSP430 and LCD
microprocessor
 Group growing to understand exact functionality of each part, as well as
the parts as a whole
August 22, 2011
 Start the implementation of the first prototype
 Start debugging the code
 Check if the code was compatible with the actual microprocessor and the
other parts
 Ask two professors this week and next to be on our group’s committee
August 29, 2011
 Continue writing/debugging code
 About half of the code should be written at this point
 Working on implementing pre-built functions into the code
September 12, 2011
 All parts should be in at this point
 Additionally, 2-3 back-up parts for each of the smaller, inexpensive units
should be present, or ordered
September 19, 2011
 Subsections of project work independently (i.e. current/voltage sensing
circuits, communication between microcontrollers, LCD showing
functionality)
 Code experiencing minimal issues
September 26, 2011
 The halfway point for the project, most of the parts should be together
 Group members brainstorm about best way to present the project
 Code to be perfected
 Full prototype works as per initial specifications
October 14, 2011
 Project continued to be tested
October 31, 2011
 All parts connected and talking to each other
 Necessary minor adjustments for the code and hardware
November 7, 2011
107



Final project was ready
Testing phase
Final paper and schematics review
November 14, 2011
 Statistical analysis of data
 Get ready for the presentation
November 21, 2011
 Final report typed and ready to present
December 1, 2011
 Group presentations this week
5.2 Future Improvements
The current project was focused heavily in the ability to give home owners a
reliable source of electricity if there was a power failure. Once this part of the
project was complete a number of other functions could be added. These
functions should be geared toward helping a user save money on their energy
bill. Since all of the applications in your home were being monitored then users
could be given a status of their current consumption.
By using this information home owners would have the ability to control their own
power usage. They would be able to set constraints on how much power they
want to use per month. Users could be alerted to see if they were on track with
their energy consumption for the month. An automated setting could even shut
down low-priority appliances in order for the bill not to exceed a certain amount
for that billing period. Meters usually sit outside of a home and home owners
were not aware of their total consumption. They may even be billed incorrectly if
they were not aware of how much power their home was consuming. The
consumption statistics for different months or years may be compared with one
another to see the history of power consumption for a home.
Currently when information about a month goes by it was lost. If all of the data
for these months could be saved than a detailed report could be generated. This
information could be available by subscription on a website. By introducing
internet communication a user would have the ability to save statistics
information and to control appliances from any remote location. Users would be
able to monitor their power usage and set schedules for operating their
appliances. This information would also be available from any mobile device. An
application for your favorite smart phone would also give the ability for a user to
have an in-home remote for their appliances.
108
Scheduling would also be an important energy saving application. If a user had a
routine schedule he or she could set their air conditioning to turn off as they leave
for work in the morning and resume functionality 1 hour before they return home.
The concept of our project was to have a house with almost every electricityconsuming appliance connected to one of our power strips. To make this a
reality though we would have to design additional power strip outlets to
accommodate many of the common receptacle configurations found in your
household. The typical wall receptacle was a type 5-15R, but other receptacle
such as the 6-15R, 10-30R, and 10-50R were commonly used for ranges, driers,
and hot water heaters. This project was design to accommodate the 5-15R
configurations and it would not work for the rest. The reason for the different
configurations was some appliances need 120 volts, 240 volts with ground and
no neutral, and also 240 volts with ground and a neutral. Each one of these
situations alters the power equations needed to calculate power consumption
and quality. This leaves the project at a great disadvantage because the biggest
energy consuming appliances use these receptacle configurations that we have
not designed to. A complete system would require more time and money and
was unfortunately not in the scope of this project.
Another improvement that could be made lies within the power strip module. Our
prototype design powers the microcontroller, sensors, and wireless
communications via battery, which we would ideally like to have an AC to DC
convertor in its place. Having a battery as the power source was inefficient and
requires more maintenance then there needs to be. When the battery fails
measurements and transmissions were stopped and the touch screen stops
recording for that power strip module until a battery was replaced. In the mean
time the user was no longer receiving an accurate monthly power cost analysis
and the appliance must be unplugged from the power strip module and plugged
back into the normal wall receptacle for it to work. This was a big inconvenience
and was probably the biggest downside currently to the project.
The programming could also be improved for more efficiency and a faster
executing time. Because all of the group members were electrical not computer
engineering majors our programming skills were probably not up to par with the
more experienced programmers out there. The more efficient your code is, the
faster the program could execute and if there was less code to process you were
also saving power consumption of the processor. Even though our code works
and it executes as it should, there would always be areas that could be improved.
Section 6: Design Summary
The last few pages of this document include a brief overview of the entire project.
The projects intentions, goals, specifications, and design were reviewed and
were presented in a compact format to give the reader the overall concept of
what the project was about. First the inspiration for the project and reasons we
109
believe it was worthy of taking up was discussed. Second the goals of the
project and what we intend to accomplish would be looked at. Third the
specifications and requirements would be reviewed. Last but not least we would
present the overall design using schematics and narratives to further clarify any
questions the ready might have.
6.1 Intentions and Goals
The intent of our project was to design a home energy management system. This
system would be versatile in that it would be beneficial to the homeowner during
both normal power operating conditions as well as during times of utility power.
Our definition of “normal power operating conditions” would be when the
homeowner was connected to a power source from the utility company. “Utility
power” was defined by the entire house running off of a generator which alone
does not provide an amount of power equal to what would be used from the
power company on a daily basis. In fact, the utility power should be an amount
that would severely limit the abilities of the homeowner to power the appliances
in their home. This was an important distinction because much of our project was
focused on the sometimes difficult choices that would need to be made during
times of power outage. Often, the average homeowner was unaware of the many
details and tremendous amount of power that go into allowing their lives to run as
normal. Questions like, “Can I run my hot water heater on a generator?”, “How
much power do I need to keep the refrigerator on?”, or “What were my limitations
for powering my computer/television/modem during power failure?” would be
simplified and answered in our design, while still giving the end-user as much
control as possible.
The main design of our project begins with the loss of normal power, at which
point the generator already connected to the user’s home would turn on in
response to a signal from a microprocessor. That microprocessor would be
constantly monitoring the current coming from the main, “normal”, power source;
as soon as the current disappears, a signal was sent to the generator to turn on
and take over power for the house. Our design would be a power strip module
that resembles the surge protector power strips which could be used at one’s
computer desk to plug in your computer and its peripherals. It would have two or
three outlets to plug in any standard utilities. These power strip modules would
have voltage and current sensing circuits which would be constantly taking
measurements related to power consumption. The power strip modules would
also contain contactor circuits to allow the user to remotely turn on and off the
appliances through the LCD touch screen monitor. This information (current and
voltage readings) would then be processed at a microcontroller (embedded in the
power strip module), and then transmitted wirelessly to the main microprocessor
in the LCD.
The wireless communications in our system would be done using a single
protocol, which was yet to be determined. Shown below in Figure 68, each
wireless device would be a transceiver (i.e. had the ability to transmit and receive
110
data); they would not communicate with each other, but only to and from the
central microprocessor (in the LCD). The wireless transceivers would be used
over a short range in our prototype; this could be extended to the end-user as
well, assuming the main processor in the end product was located near the
center of the house. This aspect would allow for lower power usage, as well as
lower costs for the transceivers.
Figure 68: Wireless communications block diagram
The LCD would contain the main “hub” microprocessor which would collect data
from the power strip modules and was where the bulk of the calculations and
decisions would be made. This display would be the interface between the user
and the software which would control whether or not outlets in the power strip
module were supplied with power. During normal operating conditions, the user
would be able to control which outlets would be turned on (supplied with power)
and also have the ability to turn off any or all of the outlets. There would also be a
section on the LCD which would provide the user with pertinent data about their
power consumption. In this section, an emphasis would be placed on “useful”
data; that is, we don’t wish to give overwhelming amounts of graphs and
numbers which may require knowledge of math to interpret and understand. The
LCD would show information in as simple a form as possible, while still providing
important, useful data. The data shown would be facts such as:





Instantaneous power consumption
Individual appliance power costs
Daily power consumption
Current rate of power consumption
Total power used since the last power bill
111
An emphasis would be placed on showing as many dollar amounts as possible,
so the user could more easily relate to the information on the screen. Abstract
figures, graphs, and even most other data would be shown for the purposes of
the demonstrations, and on a limited basis at that. This process was intended to
help the user to more accurately monitor control their power usage, in order to
meet their monthly power consumption goals.
6.2 Requirements and Specifications
The design would utilize an automatic transfer switch to transfer power to a
secondary power supply in case the primary source fails. The automatic transfer
switch was a third party device. It only needs to be tested for functionality. For
this project the use of a generator to provide secondary power in case of an
outage was intended. After power had been transferred the devices would be
able to read load from the generator and communicate this information via
wireless signal to a main microcontroller.
Other outlets in the home would also have power monitoring features. They
must have the ability to both transmit and receive instructions or data. Each
outlet would be required to read the current that was being drawn by that specific
outlet. That information would be relayed to the main microcontroller which
would sum all of the power being drawn from the home at that time. If too much
power was being drawn with respect to the generator an interrupt signal could be
sent to any of the microcontrollers that were in control of an appliance. These
microcontrollers must stop the use of the appliance immediately upon request.
Each microcontroller would measure a specific current. The accuracy of these
currents should be precise enough to prevent damage to any appliances or the
generator. We would set our initial goal to reach a precision of 1mA. The main
unit would receive power reading information from all the other microcontrollers
every second. The main unit would provide absolute instructions to all other
microcontrollers. It must have the ability to stop and resume power consumption
of any outlet in the home. These criteria would be set by the user through an
LCD input screen. In case of an overload of the secondary power source
appliances would shut down with respect to their priority.
Outlets would be sorted by priority based on user input. Each outlet priority level
would have a number ranging from 1-10 with 10 being the most important. In the
event two outlets were given the same number the user would be asked which
appliance was more important. Here the user could decide if they would rather
have warm water or air conditioning. All of the priority level inputs would be given
to the system by the user through the LCD panel. In case no priority was set the
appliance draining the largest amount of electricity would be shut down.
The power strips must be able to achieve these specifications as part of our
design:
112









Send data every 5 seconds
Enter sleep mode for 5 seconds in between packet transmissions.
Measure currents up to 15 amps
Measure voltages of 240V
Up to 2 appliances could be connected to any power strip
Have a line of sight range of at least 100 feet in an indoor environment
Have a range of at least 50 feet between two rooms
Power readings must be measured within an error margin of 5%.
Must have an ADC that was 10 bits wide for 1024 steps of resolution
The Main control station would consist of a controller that receives information
from all the power strips and displays this information on the user controlled LCD
display. This station should be able to send and receive information from the
LCD. If the LCD relays information to shut down a specific appliance an interrupt
signal must be sent to relay that information.
The specifications for this unit were as follows:





Receive data from 3 different power strips 5 seconds
Receive data from power metering device two times per second
Communicate this data to the LCD module 3 seconds
Must have an ADC that was 10 bits wide for 1024 steps of resolution
Receive and transmit information with an accuracy of 3 decimal places
with the local microcontroller.
The LCD screen was driven by its own processor and must be able to
communicate with the microcontrollers. Through the LCD screen the user should
have the ability to read power readings from any appliance in his home as well as
the readings from all of his appliances at once. The user should be able to set
priority levels for each of these outlets. In addition the user should have control
for turning on and off each outlet individually. If the user was trying to activate an
outlet that would overload the generator then they would be given a warning. The
LCD and the main control station were in constant communication.
The LCD screen would be required to do the following:

Receive and transmit information with an accuracy of 3 decimal places
with the main microcontroller.
While an outlet was not being used it should enter a sleep mode to preserve the
battery life of that specific unit. The microcontroller would remain in sleep mode
as long as there was no signal telling it to turn on. Once the microcontroller
receives the signal to turn on it would return to normal operation.
The last power metering device was the one reading the power information being
delivered by the generator. This device relays the information to the main
microcontroller. All of the overloading decisions of this project were dependent
113
upon this part. This device must be able to read the power load on a generator
by reading voltage and current information.
The generator power metering device was required to do the following:





Read power supply information from the generator twice a second
Send the power consumption information to the main microcontroller twice
a second
Must have an accuracy of 3 decimal places
Measure currents up to 30 amps
Measure voltages of 240V
To begin we start off with Figure 69 shown below and this was the overall block
diagram view of the entire project. As you could see the project consists of four
main components including the ATS module, Power strip module, Generator
module, and the user display. Each module would be briefly discussed in the
proceeding paragraphs.
114
Figure 69: Project Block Diagram
6.3 Power Strip Module
The power strip module would contain all of the components we researched and
documented in our design. The main components of the power strip module
include a duplex receptacle, power relay for switching the receptacle on and off,
a current transformer for measuring current, a potential transformer for
measuring voltage, a MSP430 microcontroller for power calculation and last but
not least a RF wireless transceiver for sending a receiving data from the main
touch screen monitor. Figure 70 below was the wiring schematic of the power
strip module and give a clear indication of the functionality of the circuit. Please
115
refer back to the design section for a breakout of all the components and their
functions.
Figure 70: Power Strip Module Schematic
6.4 ATS Module
The automatic transfer switch must have the capability to wirelessly transmit
whether it was in normal or emergency mode back to the touch screen controller.
The transfer switch already provides a means of signaling whether it was utilizing
normal or emergency power via low voltage signal so all we have to do was send
that signal back to the touch screen controller. We were going to use the same
TI CC2500 wireless chip that would be utilized in the power strip modules and
touch screen so adding it to the network would be without much effort. The
rating of the CC2550 interrupt input rating was 3.3 volts and the transfer switch’s
116
signaling circuit outputs 12 volts so a voltage dividing circuit would need to be
added to make the two compatible. In addition the CC2500 requires a DC power
supply between 1.8 and 3.6 volts and we would utilize a battery to supply this
voltage. Because of the ease of removing the service compartment cover of the
transfer switch and access to terminals 5 and 6 we believe that making a small
module with battery power would be the smallest and most economical option to
us. It would be extremely easy for a homeowner to change the battery when it
was time to be replaced. Below in figure 71 a simple block diagram of the
automatic transfer switch module was shown.
UTILITY
POWER
HOUSE
PANEL
ATS
GENERATOR
POWER
VOLTAGE
DIVIDING
CIRCUIT
CC2500
WIRELESS
1.8 -3.6 DC
POWER
SUPPLY
QSCREEN
CONTROLLER
Figure 71: ATS Module Block Diagram
6.5 Generator Module
During times of power loss the home generator becomes the lifeline of the homes
power system. In order to ensure the safeguarding of the generators operating
conditions we must design a module to measure the generators instantaneous
power consumption and have that information wirelessly transmitted back to the
touch screen controller for program manipulations. The output from a typical
home generator was 240 volts with a neutral. The ground was picked up from
the houses existing grounding system. The power measurement module for the
generator would be slightly different than that of the power strip modules
117
because we have two phases to monitor instead of one. This was not a problem
though as we only need one additional power measuring circuit to accommodate
for the extra phase. The total power of the generator was simply the addition of
the two individual phases. Except with the addition of the extra power measuring
circuit the main components remain the same as the power strip modules. First
the power would be measured via current and voltage sensors and then power
calculations would be done using the MSP 430. This information would then be
wirelessly transmitted back to the touch screen controller for further processing
and display to the user. A block diagram of the generator module could be seen
below in figure 72.
GENERATOR
PHASE
A
PHASE
B
POWER
MEASURING
CICUIT
POWER
MEASURING
CICUIT
QSREEN
CONTROLLER
BATTERY
POWER
CC2500
WIRELESS
MSP 430
Figure 72: Generator Module Block Diagram
6.7 User interface
After careful considerations of the needs and goals of this project we found that
the QScreen Controller tm from Mosaic Industries, Inc. best suited those
requirements.
The QScreen Controller tm was a combination of a Cprogrammable single-board computer with a touch screen operated graphical
user interface. This touch screen monitor was powered by the Motorola 68hc11
microcontroller, which we were very familiar with from the laboratory section of
Embedded Systems class. The fact the each of the group members was familiar
118
with the C-programming language as well as the assembly language of the
Motorola 68hc11 was what lead us to selecting this product for our project.
Because this project was a proto-type of product that would be utilized in
consumer household, appearance and size were of great importance and
concern. We envision this touch screen controller to be mounted near the
houses existing thermostat so for cosmetic purposes we want our display to be
about the same size as a standard thermostat. After taking measurements of a
few thermostats we found the average to be 4” tall and 6” wide. The QScreen
controllers overall dimensions were 4.125” tall and 6” wide making it the ideal
candidate for size requirements. A picture of the overall appearance and design
of the QScreen controller could be seen below in Figure 73. As you could tell
from the figure below the QScreen controller would surface mount perfect in a
household wall with a cutout not much bigger then needed for a typical switch or
receptacle outlet box.
Figure 73: QScreen Controller
Permission requested from Moasic Industries
The QScreen controller was equipped with a touchscreen operated graphical
user interface on a high-contrast 128X240 pixel display with a 5X4 touchscreen
overlay. The screen itself measures 4.8” diagonally and was a high contrast
CCFL white-on-blue monochrome LCD. The user does have the ability to control
backlight and contrast of the screen using software provided with the product.
Because of the ability to navigate through different menus this size screen was
more than enough to satisfy the requirements of this project.
As noted earlier the main reason for using the QScreen controller was having
previous knowledge of the Motorola 68hc11 as well as the C programming
language.
The QScreen controller gives the programmer the option of
programming in either Control C or QED-Forth. We chose to use Control C as
our programming language for the programming of this project. Control C tm
119
cross-compiler comes with the purchase of the QScreen controller was written by
Fabius Software Systems and customized by Mosaic Industries to facilitate
programming the QScreen controller in C. It was a full ANSI C compiler and
macro pre-processor. The compiler supports floating-point math, structures and
unions, and allows you to program familiar C syntax. Programming was easy
done using an interactive debugger and multitasking executive. The controller
was pre-loaded with hardware control routines including drawing functions for the
display. After programming in ANSI C by compiling our code on our PC and
downloading the code to the QScreen Controller it would then be automatically
executed. The real time operating system and onboard flash memory manages
all required initializations and auto-starts our application code.
Now that we have selected the touch screen monitor for the project we need to
design the functionality and appear of the interface to the user. We want to avoid
a busy screen and keep it simple and easy for the user to read. Multiple pages
would be avoided so as not to lose the user in navigation throughout the
program. A preliminary touch screen output could be seen below in figure 74.
As you could see the legend indicated that a boxed area with the dotted
background would be the only input area for the user. These areas include the
adding a module, turning modules on and off, and also the ability to input the
power rating of their home generator.
ADD MODULE
MODULE #
GENERATOR
KW RATING
ON/OFF
CURREN
T KW
KW
AVERAG
EKW
DAILY
COST
MONTHLY
COST
1
2
CURRENT GENERATOR CONSUMPTION
KW
REMAINING GENERATOR POWER
KW
Figure 74: LCD Graphical Interface
We want to make sure the product was expandable to the user, which was why
we included the “Add Module” feature. This allows the user to make the initial
investment of the touch screen controller and maybe one or two power strip
modules to test out and use in their house. At any time in the future they could
always purchase another power strip module and connect it to their existing
120
network. All the user need to do was plug in the power strip module wherever
they would like to use it and then simply walk to the touch screen monitor and
click “Add Module”. This would automatically recognize the power strip module
just plugged in and would assign to it the next number in the module list. It was
important to note that if the user wished to add multiple power strip modules at
the same time that they would have to plug each one in separately and click the
“Add Module” button before plugging in the next. The reason for this was when
the user clicks “Add Module” the program was going to be looking for new
devices and if more than one was available an error might occur.
The next feature the user would be able to control was turning the power strip
modules on and off. This option could be used during normal power operations,
but was mainly intended to be used during times of emergency generator power.
During normal power operation the user should have all the modules powered on
and simply turn off their appliance how they normally do. The only advantage to
using the on and off feature during this time was to conserve power on appliance
that draw power even when they were off such as televisions, computer, etc.
This way you could shut the power off to them via the touch screen without
having to unplug each power cord from the wall. The main purpose of the on and
off feature was used when the system was in emergency mode. As soon as
normal power was lost and the transfer switch switches over to emergency mode
the QScreen controller would automatically shut off all power strip modules and
wait for the user to decide what modules they want on. By looking at the
“average power consumption” column the user could identify normal
instantaneous power draw from individual appliances measured in KW. The
QScreen knows the KW rating of the generator since the user initially input its
value. As the user turns on individual power strip modules the program would
calculate the remaining available generator power and report it back to the
screen. This way the user would be able realize when they were approaching
the limitations of the generator. Safety features would be added to insure the
user does not accidentally turn on an appliance that would overload the
generator causing failure.
The user would also have the ability to enter the KW rating of their home
generator. This was important for the “emergency mode” operation of project.
The software would monitor the usage of power from the generator as well as all
the power strip modules and would compare to make sure the power
consumption of the all the power strip modules together does not exceed the
rating of the generator. If this were to happen the generator would fail and the
project would have not satisfied its goal.
Section 7: Conclusion
The engineering industry had been built around the needs of the consumer, be it
those of individuals, companies, or military corporations. Necessity was the
mother of invention (perhaps in our field more than any other), and everything
121
that goes into our jobs and products (what we do for the products we sell, how
we do it, and what features go into it) was almost always derived from the enduser’s needs and means.
Through the last decade, our country had seen enough financial hardship and
stress to suffice for a lifetime, and as a result, more and more people were
concerned over every penny that comes and goes through their households.
More recently, this had encouraged the practice of trimming the use and overuse
of household utilities to save money on everything from gas to water and
electricity. While their efforts were certainly applauded, most homeowners were
not aware of the numerous appliances that they take for granted which may drain
a large portion of their monthly energy budgets. Turning off the A/C in the
summer or the heat in the winter was one option that many have embraced, but
having the ability to monitor all of the smaller appliances in the home may also
help lead to more energy-aware consumers. Thus, one of the goals of our project
was to aid the homeowner in his goal of lowering his monthly electric bill by
showing the power usage as well as the costs of running appliances.
Another facet of our project for a home energy monitoring system was an
automatic transfer switch (ATS) from normal (electric company) to utility
(generator) power, a feature which had been neglected by many of the
competing models. Not only was this interesting on its own, but was an extremely
helpful tool to those affected by hurricanes and other strong thunderstorms each
year in Central Florida. The ATS in our design would ensure that not a minute
would go by where those caught in storms would not have access to the many
things they may need, including: access to the media (television, radio, internet,
etc) with potentially life-saving information about the surrounding weather; the
ability to charge wireless electronic devices (cell/smart phones, tablets) with
satellite access; a running refrigerator to keep food fresh while the power lines
were down; and a hot water heater to take showers, clean clothes, and wash
dishes. In addition, our design does not allow for an overload of the generator,
helping to save the user the cost of replacing it.
Technology to conserve energy was certainly not new, but it was at an early
stage in its development; it had a long way to go, and many innovations along
the way. Nor was the concept to monitor utility usage new or novel. In fact, most
of the large companies that design and manufacture microprocessors have had
whole series of products specifically targeted toward monitoring and measuring
utility usage, and peripherals associated with it. However, our project combines
these ideas, along with an automatic transfer switch, into a home energy system
that had the potential to stand out among and even rise above its competitors.
Section 8: Product Manual
The user manual for our product would be as simple as possible, and ideally
should only consist of two or three different commands. The process for running
122
and using our system was detailed below and was in two parts: the first part was
during normal power operations, and the second part was during a power failure.
Normal operating conditions (main power)
1.
2.
in
To begin, wake the system up by touching anywhere on the screen, or by
pressing any of the buttons on the side.
You would be greeted with a home screen similar to the one shown below
Figure 75. Each port shown displays power currently consumed as well
as its current state (on/off).
Figure 75: LCD home screen
3.
or
port’s
4.
5.
6.
To switch the state of an outlet/utility (i.e. to turn on outlet on that was off,
to turn an outlet off that was on), simply press the button beneath that
name, labeled “TURN ON/OFF”.
To turn all ports on, press the button “ALL ON”, located near the bottom of
your screen.
Conversely, to turn all ports off, press the button “ALL OFF”, also located
near the bottom of your screen.
At the top of the screen would be a running total of KWh for the month, as
well as a percentage, which was the percentage of your monthly goal you
have used thus far. By clicking the button labeled “MANAGE” beneath this
123
total, a separate screen would appear which would allow you to change
this goal.
7.
On the new screen, the user would see their current monthly goal for
power
consumption. By pressing the button next to this screen
labeled “EDIT”,
the user could input any goal he/she wishes, and the
program would
automatically update with this information.
8.
Next, the user could select “EXIT” on this screen, and then “MAIN
SCREEN” on the next screen, and would be returned to the first screen.
Utility (generator) operating conditions:
1.
During utility operating conditions, the user would be greeted with the
same screen as before (see Figure 400). Each port shown displays power
currently consumed as well as its current state (on/off). Each port would
also display a percentage number which corresponds to the percentage of
power that unit consumes with respect to the amount of power available
from the generator.
2.
To switch the state of an outlet/utility (i.e. to turn on outlet on that was off,
or
to turn an outlet off that was on), simply press the button beneath that
port’s name, labeled “TURN ON/OFF”.
3.
To turn all ports on, press the button “ALL ON”, located near the bottom of
your screen. However, if doing so would overload the generator, the
screen
will read “WARNING: GENERATOR CAPACITY REACHED”, and
would not
change the state of any of the connected ports.
4.
Conversely, to turn all ports off, press the button “ALL OFF”, also located
near the bottom of your screen.
5.
When normal power was restored, the main home screen would return to
how it looked in the first part of the manual, and the user would no longer be
limited
to using a certain number of ports.
124